2BFT — 2 Boys From Town
2BFTVaniyambadi, IN
The catalog · July 2026Updated weekly

100+ agentic systems
you can build today.

Real workflows, real tools, real impact. From beginner to advanced. All the systems 2BFT uses, teaches, and ships. Copy, customize, ship.

105
systems
10
categories
5
patterns
Free
all tools
Browse by category →Want us to build one? Mail us →
10 categories

Pick the world you live in. Pick a system that ships this week.

Every category below has 5–15 real systems, each with a name, what it does, the tools to build it, and a difficulty badge. Click any system to jump to the full catalog.
01 / 10

Personal Productivity

Your day, automated. Wake up, ship, sleep.

15 systems→
02 / 10

Coding & Development

Ship code without typing code. AI as your pair-programmer.

15 systems→
03 / 10

Content & Creative

Write, design, ship — at the pace the algorithm wants.

15 systems→
04 / 10

Business & Marketing

Sales, support, and growth on autopilot. Boring repeatable wins.

15 systems→
05 / 10

WhatsApp & Instagram

Where your customers already are. Reply before they tap away.

10 systems→
06 / 10

E-commerce & D2C

Shopify, D2C, marketplace. The agent does the boring work.

10 systems→
07 / 10

Research & Learning

Read 10x faster. Learn what matters. Skip the rest.

10 systems→
08 / 10

Health & Lifestyle

Habits, meals, sleep. Boring stuff, automated so you actually do it.

5 systems→
09 / 10

Real Estate & Local Business

Local shops, brokers, clinics. The agent does the follow-up.

5 systems→
10 / 10

Education & Students

Study smarter, not longer. For students and lifelong learners.

5 systems→
Category 0115 systems

Personal Productivity

Your day, automated. Wake up, ship, sleep.

01

Daily AI morning briefing

Beginner

Overnight news, weather, calendar, top 3 priorities — pushed to Telegram before you wake up.

ClaudeTelegram MCPTavily
⧉ Brief
Goal

Overnight news, weather, calendar, top 3 priorities — pushed to Telegram before you wake up.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Overnight news, weather, calendar, top 3 priorities — pushed to Telegram before you wake up.
Starter prompt
Build an agent called "Daily AI morning briefing".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Daily AI morning briefing · 3 tools
02

Email-to-task extractor

Beginner

Gmail inbox → Notion tasks, with priority, owner, and due date pulled from the email body.

GPT-5.6Gmail APINotion API
⧉ Brief
Goal

Gmail inbox → Notion tasks, with priority, owner, and due date pulled from the email body.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Gmail inbox → Notion tasks, with priority, owner, and due date pulled from the email body.
Starter prompt
Build an agent called "Email-to-task extractor".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Email-to-task extractor · 3 tools
03

Calendar conflict detector

Beginner

Watches Google Calendar, flags double-bookings and back-to-back meetings with no buffer.

ClaudeCalendar MCPMake
⧉ Brief
Goal

Watches Google Calendar, flags double-bookings and back-to-back meetings with no buffer.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Watches Google Calendar, flags double-bookings and back-to-back meetings with no buffer.
Starter prompt
Build an agent called "Calendar conflict detector".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Calendar conflict detector · 3 tools
04

Meeting notes → action items

Beginner

Records meeting audio, transcribes, extracts action items, and pushes them to Trello with owner tags.

WhisperGPT-5.6Trello API
⧉ Brief
Goal

Records meeting audio, transcribes, extracts action items, and pushes them to Trello with owner tags.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Records meeting audio, transcribes, extracts action items, and pushes them to Trello with owner tags.
Starter prompt
Build an agent called "Meeting notes → action items".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Meeting notes → action items · 3 tools
05

Daily journaling companion

Beginner

End-of-day prompt in Notion → Claude replies with patterns, wins, and one question for tomorrow.

ClaudeNotion MCPCron
⧉ Brief
Goal

End-of-day prompt in Notion → Claude replies with patterns, wins, and one question for tomorrow.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →End-of-day prompt in Notion → Claude replies with patterns, wins, and one question for tomorrow.
Starter prompt
Build an agent called "Daily journaling companion".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Daily journaling companion · 3 tools
06

Habit tracker with weekly insights

Beginner

Logs habits to a Notion database, sends a Sunday summary with streak, drops, and one nudge.

ClaudeNotionTelegram MCP
⧉ Brief
Goal

Logs habits to a Notion database, sends a Sunday summary with streak, drops, and one nudge.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Logs habits to a Notion database, sends a Sunday summary with streak, drops, and one nudge.
Starter prompt
Build an agent called "Habit tracker with weekly insights".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Habit tracker with weekly insights · 3 tools
07

Recipe generator from fridge photo

Beginner

Snap your fridge → Claude names the dish, lists ingredients, and gives 5 steps.

Nano Banana 2ClaudeNotion
⧉ Brief
Goal

Snap your fridge → Claude names the dish, lists ingredients, and gives 5 steps.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Snap your fridge → Claude names the dish, lists ingredients, and gives 5 steps.
Starter prompt
Build an agent called "Recipe generator from fridge photo".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Recipe generator from fridge photo · 3 tools
08

Travel planner

Beginner

Dates + budget + preferences → day-by-day itinerary with maps, food, and rain-day backups.

GPT-5.6MakeTavily
⧉ Brief
Goal

Dates + budget + preferences → day-by-day itinerary with maps, food, and rain-day backups.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Dates + budget + preferences → day-by-day itinerary with maps, food, and rain-day backups.
Starter prompt
Build an agent called "Travel planner".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Travel planner · 3 tools
09

Resume-to-cover-letter

Intermediate

Drop a PDF resume + job description → personalized cover letter, edited to your voice.

ClaudeOpenClawPDF parser
⧉ Brief
Goal

Drop a PDF resume + job description → personalized cover letter, edited to your voice.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Drop a PDF resume + job description → personalized cover letter, edited to your voice.
Starter prompt
Build an agent called "Resume-to-cover-letter".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Resume-to-cover-letter · 3 tools
10

Inbox zero assistant

Intermediate

Reads unread email, drafts replies in your voice, surfaces only the ones that need you.

ClaudeGmail MCPOpenClaw
⧉ Brief
Goal

Reads unread email, drafts replies in your voice, surfaces only the ones that need you.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Reads unread email, drafts replies in your voice, surfaces only the ones that need you.
Starter prompt
Build an agent called "Inbox zero assistant".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Inbox zero assistant · 3 tools
11

Family chore scheduler

Intermediate

WhatsApp group → Claude parses requests → slots chores into shared Google Calendar.

OpenClawWhatsApp APICalendar MCP
⧉ Brief
Goal

WhatsApp group → Claude parses requests → slots chores into shared Google Calendar.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →WhatsApp group → Claude parses requests → slots chores into shared Google Calendar.
Starter prompt
Build an agent called "Family chore scheduler".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Family chore scheduler · 3 tools
12

Bill tracker

Intermediate

Photo of a receipt → expense line item, categorized, with monthly totals in a sheet.

Nano Banana 2OpenClawSheets
⧉ Brief
Goal

Photo of a receipt → expense line item, categorized, with monthly totals in a sheet.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Photo of a receipt → expense line item, categorized, with monthly totals in a sheet.
Starter prompt
Build an agent called "Bill tracker".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Bill tracker · 3 tools
13

Voice journal → themes

Intermediate

Record 2 minutes a day, Claude finds recurring themes, blockers, and wins across the week.

ElevenLabsClaudeNotion
⧉ Brief
Goal

Record 2 minutes a day, Claude finds recurring themes, blockers, and wins across the week.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Record 2 minutes a day, Claude finds recurring themes, blockers, and wins across the week.
Starter prompt
Build an agent called "Voice journal → themes".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Voice journal → themes · 3 tools
14

Learning path generator

Intermediate

Goal + current level → 8-week curriculum with daily tasks, resources, and checkpoints.

ClaudeTavilyNotebookLM
⧉ Brief
Goal

Goal + current level → 8-week curriculum with daily tasks, resources, and checkpoints.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Goal + current level → 8-week curriculum with daily tasks, resources, and checkpoints.
Starter prompt
Build an agent called "Learning path generator".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Learning path generator · 3 tools
15

Daily standup bot

Intermediate

Slack prompts the team at 9:55, collates answers, posts a standup summary to the channel.

ClaudeSlack MCPCron
⧉ Brief
Goal

Slack prompts the team at 9:55, collates answers, posts a standup summary to the channel.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Slack prompts the team at 9:55, collates answers, posts a standup summary to the channel.
Starter prompt
Build an agent called "Daily standup bot".
Trigger: <TIME/FREQUENCY>.
Read from: <SOURCE>.
Write to: <DESTINATION>.
Don't ask me for confirmation between steps. Run it.
If something fails, retry once, then log and skip.
Daily standup bot · 3 tools
Category 0215 systems

Coding & Development

Ship code without typing code. AI as your pair-programmer.

16

Code review bot

Intermediate

PR opens → Claude Code reads the diff, posts review comments, flags risks, and suggests tests.

Claude CodeGitHub MCPMastra
⧉ Brief
Goal

PR opens → Claude Code reads the diff, posts review comments, flags risks, and suggests tests.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →PR opens → Claude Code reads the diff, posts review comments, flags risks, and suggests tests.
Starter prompt
Build the "Code review bot" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Code review bot · 3 tools
17

Documentation updater

Intermediate

Detects code changes that aren't reflected in docs, drafts the missing section, opens a PR.

Claude CodeMkDocsGitHub
⧉ Brief
Goal

Detects code changes that aren't reflected in docs, drafts the missing section, opens a PR.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Detects code changes that aren't reflected in docs, drafts the missing section, opens a PR.
Starter prompt
Build the "Documentation updater" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Documentation updater · 3 tools
18

Test generator

Intermediate

Drop a function → unit tests covering happy path, edge cases, and assertions. Jest, pytest, or vitest.

Claude Sonnet 5JestAider
⧉ Brief
Goal

Drop a function → unit tests covering happy path, edge cases, and assertions. Jest, pytest, or vitest.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Drop a function → unit tests covering happy path, edge cases, and assertions. Jest, pytest, or vitest.
Starter prompt
Build the "Test generator" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Test generator · 3 tools
19

Bug triage

Intermediate

New issue arrives → Claude scores priority, suggests an owner, asks the right clarifying question.

ClaudeLinear MCPGitHub
⧉ Brief
Goal

New issue arrives → Claude scores priority, suggests an owner, asks the right clarifying question.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →New issue arrives → Claude scores priority, suggests an owner, asks the right clarifying question.
Starter prompt
Build the "Bug triage" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Bug triage · 3 tools
20

Documentation Q&A

Intermediate

RAG over your internal docs — answers with citations, refuses when the answer isn't there.

RAGFlowClaudepgvector
⧉ Brief
Goal

RAG over your internal docs — answers with citations, refuses when the answer isn't there.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →RAG over your internal docs — answers with citations, refuses when the answer isn't there.
Starter prompt
Build the "Documentation Q&A" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Documentation Q&A · 3 tools
21

PR description writer

Beginner

Diff → human-readable PR title, summary, test plan, and screenshot placeholders.

Claude CodeGitHubMastra
⧉ Brief
Goal

Diff → human-readable PR title, summary, test plan, and screenshot placeholders.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Diff → human-readable PR title, summary, test plan, and screenshot placeholders.
Starter prompt
Build the "PR description writer" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
PR description writer · 3 tools
22

TypeScript migration assistant

Advanced

JS file in, TypeScript file out — types inferred, any-es kept, build still passes.

Claude Sonnet 5Babelts-migrate
⧉ Brief
Goal

JS file in, TypeScript file out — types inferred, any-es kept, build still passes.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →JS file in, TypeScript file out — types inferred, any-es kept, build still passes.
Starter prompt
Build the "TypeScript migration assistant" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
TypeScript migration assistant · 3 tools
23

API endpoint monitor

Advanced

Watches API logs, detects error spikes, posts a Mastra alert with the failing request and stack.

ClaudeMastraPagerDuty
⧉ Brief
Goal

Watches API logs, detects error spikes, posts a Mastra alert with the failing request and stack.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Watches API logs, detects error spikes, posts a Mastra alert with the failing request and stack.
Starter prompt
Build the "API endpoint monitor" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
API endpoint monitor · 3 tools
24

Database query helper

Intermediate

Natural language question → safe SQL with table joins, indexes used, and a one-line explanation.

ClaudePostgres MCPMastra
⧉ Brief
Goal

Natural language question → safe SQL with table joins, indexes used, and a one-line explanation.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Natural language question → safe SQL with table joins, indexes used, and a one-line explanation.
Starter prompt
Build the "Database query helper" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Database query helper · 3 tools
25

Refactor agent

Advanced

Reads a file, suggests cleaner structure, applies the diff, and re-runs the test suite.

Claude Sonnet 5AiderJest
⧉ Brief
Goal

Reads a file, suggests cleaner structure, applies the diff, and re-runs the test suite.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Reads a file, suggests cleaner structure, applies the diff, and re-runs the test suite.
Starter prompt
Build the "Refactor agent" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Refactor agent · 3 tools
26

Vulnerability scanner

Advanced

Runs Semgrep, reads findings, drafts a security report ranked by exploitability.

OpenClawSemgrepClaude
⧉ Brief
Goal

Runs Semgrep, reads findings, drafts a security report ranked by exploitability.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Runs Semgrep, reads findings, drafts a security report ranked by exploitability.
Starter prompt
Build the "Vulnerability scanner" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Vulnerability scanner · 3 tools
27

Spec-to-code

Advanced

Paste a PRD → Claude Code scaffolds the project, writes the first feature, and opens a PR.

Claude CodeContinue.devGitHub
⧉ Brief
Goal

Paste a PRD → Claude Code scaffolds the project, writes the first feature, and opens a PR.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Paste a PRD → Claude Code scaffolds the project, writes the first feature, and opens a PR.
Starter prompt
Build the "Spec-to-code" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Spec-to-code · 3 tools
28

Git history summarizer

Intermediate

Last 50 commits → human changelog grouped by feature, with breaking changes called out.

ClaudeMastraGitHub
⧉ Brief
Goal

Last 50 commits → human changelog grouped by feature, with breaking changes called out.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Last 50 commits → human changelog grouped by feature, with breaking changes called out.
Starter prompt
Build the "Git history summarizer" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Git history summarizer · 3 tools
29

Dependency update explainer

Beginner

package.json changes → plain-English release notes for each new version, with risk flags.

ClaudenpmMastra
⧉ Brief
Goal

package.json changes → plain-English release notes for each new version, with risk flags.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →package.json changes → plain-English release notes for each new version, with risk flags.
Starter prompt
Build the "Dependency update explainer" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
Dependency update explainer · 3 tools
30

README generator

Beginner

Empty repo → Claude scans the tree, writes a README with install, run, and contribution steps.

Claude CodeGitHub
⧉ Brief
Goal

Empty repo → Claude scans the tree, writes a README with install, run, and contribution steps.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Empty repo → Claude scans the tree, writes a README with install, run, and contribution steps.
Starter prompt
Build the "README generator" system in my repo.
1. Read the relevant files (don't guess).
2. Plan the change in 5 bullets before writing code.
3. Apply the diff.
4. Run the test suite. If anything fails, fix and re-run — don't stop.
5. Open a PR with a clear title and description.
README generator · 2 tools
Category 0315 systems

Content & Creative

Write, design, ship — at the pace the algorithm wants.

31

Instagram post generator

Beginner

Product photo + 5-word prompt → caption, hashtags, and 4 carousel slides in your brand voice.

ClaudeNano Banana 2OpenClaw
⧉ Brief
Goal

Product photo + 5-word prompt → caption, hashtags, and 4 carousel slides in your brand voice.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Product photo + 5-word prompt → caption, hashtags, and 4 carousel slides in your brand voice.
Starter prompt
Create "Instagram post generator" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Instagram post generator · 3 tools
32

YouTube script writer

Intermediate

Topic + audience + length → hook, intro, 5 chapters, B-roll cues, and CTA in your voice.

Claude Sonnet 5NotebookLMMake
⧉ Brief
Goal

Topic + audience + length → hook, intro, 5 chapters, B-roll cues, and CTA in your voice.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Topic + audience + length → hook, intro, 5 chapters, B-roll cues, and CTA in your voice.
Starter prompt
Create "YouTube script writer" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
YouTube script writer · 3 tools
33

Reel storyboard generator

Intermediate

Topic → 10-frame storyboard with shot type, on-screen text, and Veo 3.1 prompts per frame.

Nano Banana 2ClaudeVeo 3.1
⧉ Brief
Goal

Topic → 10-frame storyboard with shot type, on-screen text, and Veo 3.1 prompts per frame.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Topic → 10-frame storyboard with shot type, on-screen text, and Veo 3.1 prompts per frame.
Starter prompt
Create "Reel storyboard generator" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Reel storyboard generator · 3 tools
34

Podcast show notes generator

Intermediate

Audio file → transcript, 5 key quotes, chapter markers, and 3 tweet drafts.

WhisperClaudeMastra
⧉ Brief
Goal

Audio file → transcript, 5 key quotes, chapter markers, and 3 tweet drafts.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Audio file → transcript, 5 key quotes, chapter markers, and 3 tweet drafts.
Starter prompt
Create "Podcast show notes generator" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Podcast show notes generator · 3 tools
35

Tweet thread writer

Beginner

One-line idea → 7-tweet thread with hook, narrative arc, and a soft CTA.

GPT-5.6ClaudeMake
⧉ Brief
Goal

One-line idea → 7-tweet thread with hook, narrative arc, and a soft CTA.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →One-line idea → 7-tweet thread with hook, narrative arc, and a soft CTA.
Starter prompt
Create "Tweet thread writer" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Tweet thread writer · 3 tools
36

Email newsletter writer

Beginner

Topic + audience → 8-section newsletter with subject, preview, intro, 5 ideas, and PS.

Claude Sonnet 5Mailchimp
⧉ Brief
Goal

Topic + audience → 8-section newsletter with subject, preview, intro, 5 ideas, and PS.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Topic + audience → 8-section newsletter with subject, preview, intro, 5 ideas, and PS.
Starter prompt
Create "Email newsletter writer" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Email newsletter writer · 2 tools
37

Blog post + image bundle

Intermediate

Topic → 1,200-word post + hero image + 3 in-article images in your brand style.

ClaudeNano Banana 2Notion
⧉ Brief
Goal

Topic → 1,200-word post + hero image + 3 in-article images in your brand style.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Topic → 1,200-word post + hero image + 3 in-article images in your brand style.
Starter prompt
Create "Blog post + image bundle" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Blog post + image bundle · 3 tools
38

Repurpose content

Intermediate

Long blog → 5 platform-specific posts: Twitter thread, LinkedIn, IG carousel, Reel script, newsletter.

Claude Sonnet 5OpenClaw
⧉ Brief
Goal

Long blog → 5 platform-specific posts: Twitter thread, LinkedIn, IG carousel, Reel script, newsletter.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Long blog → 5 platform-specific posts: Twitter thread, LinkedIn, IG carousel, Reel script, newsletter.
Starter prompt
Create "Repurpose content" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Repurpose content · 2 tools
39

SEO content brief

Intermediate

Keyword → full brief: search intent, top 10 SERPs, outline, word count, and FAQs to answer.

ClaudePerplexityAhrefs
⧉ Brief
Goal

Keyword → full brief: search intent, top 10 SERPs, outline, word count, and FAQs to answer.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Keyword → full brief: search intent, top 10 SERPs, outline, word count, and FAQs to answer.
Starter prompt
Create "SEO content brief" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
SEO content brief · 3 tools
40

Voice-over generator

Beginner

Script + voice profile → 5-minute audio file with pauses, emphasis, and breath marks.

ElevenLabs v3Claude
⧉ Brief
Goal

Script + voice profile → 5-minute audio file with pauses, emphasis, and breath marks.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Script + voice profile → 5-minute audio file with pauses, emphasis, and breath marks.
Starter prompt
Create "Voice-over generator" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Voice-over generator · 2 tools
41

Music for content

Beginner

Mood + length → royalty-free music track, no vocals, with stems if you need edits.

SunoClaudeMake
⧉ Brief
Goal

Mood + length → royalty-free music track, no vocals, with stems if you need edits.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Mood + length → royalty-free music track, no vocals, with stems if you need edits.
Starter prompt
Create "Music for content" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Music for content · 3 tools
42

Video shorts from long video

Advanced

10-minute video → 5 shorts under 60 seconds, each with its own hook and end-CTA.

Veo 3.1ClaudeMastra
⧉ Brief
Goal

10-minute video → 5 shorts under 60 seconds, each with its own hook and end-CTA.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →10-minute video → 5 shorts under 60 seconds, each with its own hook and end-CTA.
Starter prompt
Create "Video shorts from long video" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Video shorts from long video · 3 tools
43

Quote cards from articles

Beginner

Article URL → 6 quote cards, each under 240 chars, with a designed graphic in your palette.

Nano Banana 2ClaudeOpenClaw
⧉ Brief
Goal

Article URL → 6 quote cards, each under 240 chars, with a designed graphic in your palette.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Article URL → 6 quote cards, each under 240 chars, with a designed graphic in your palette.
Starter prompt
Create "Quote cards from articles" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Quote cards from articles · 3 tools
44

Carousel generator

Intermediate

Long article → 10-slide IG carousel with one idea per slide, designed in your template.

Nano Banana 2ClaudeMake
⧉ Brief
Goal

Long article → 10-slide IG carousel with one idea per slide, designed in your template.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Long article → 10-slide IG carousel with one idea per slide, designed in your template.
Starter prompt
Create "Carousel generator" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Carousel generator · 3 tools
45

Newsletter digest

Intermediate

20 saved articles → 1 weekly digest grouped by theme, with your hot takes layered in.

ClaudeTavilyMailchimp
⧉ Brief
Goal

20 saved articles → 1 weekly digest grouped by theme, with your hot takes layered in.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →20 saved articles → 1 weekly digest grouped by theme, with your hot takes layered in.
Starter prompt
Create "Newsletter digest" for the following inputs.
Topic: <TOPIC>
Audience: <WHO>
Voice: <TONE> (or use the brand voice in /brand-voice.md if it exists)
Length: <WORDS/MINUTES>
Output in the same format every time. Don't add explanations before or after.
Newsletter digest · 3 tools
Category 0415 systems

Business & Marketing

Sales, support, and growth on autopilot. Boring repeatable wins.

46

Lead qualifier

Intermediate

Form fill → lead score, segment, and a CRM tag — sales team only sees MQL and above.

ClaudeOpenClawHubSpot
⧉ Brief
Goal

Form fill → lead score, segment, and a CRM tag — sales team only sees MQL and above.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Form fill → lead score, segment, and a CRM tag — sales team only sees MQL and above.
Starter prompt
Build "Lead qualifier" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Lead qualifier · 3 tools
47

Customer support bot

Intermediate

FAQ + your docs + Claude → instant answers, escalates to a human when the question is fuzzy.

ClaudeOpenClawIntercom
⧉ Brief
Goal

FAQ + your docs + Claude → instant answers, escalates to a human when the question is fuzzy.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →FAQ + your docs + Claude → instant answers, escalates to a human when the question is fuzzy.
Starter prompt
Build "Customer support bot" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Customer support bot · 3 tools
48

Onboarding email sequence

Beginner

Signup event → 5 emails over 10 days, with personalized tips based on user behaviour.

Claude Sonnet 5MailchimpOpenClaw
⧉ Brief
Goal

Signup event → 5 emails over 10 days, with personalized tips based on user behaviour.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Signup event → 5 emails over 10 days, with personalized tips based on user behaviour.
Starter prompt
Build "Onboarding email sequence" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Onboarding email sequence · 3 tools
49

Win-back campaign

Advanced

Churned user → 3-touchpoint sequence: value reminder, new feature, soft discount offer.

ClaudeOpenClawMailchimp
⧉ Brief
Goal

Churned user → 3-touchpoint sequence: value reminder, new feature, soft discount offer.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Churned user → 3-touchpoint sequence: value reminder, new feature, soft discount offer.
Starter prompt
Build "Win-back campaign" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Win-back campaign · 3 tools
50

Pricing page generator

Intermediate

Competitor URLs → 3-tier pricing with feature diff and your positioning, ready to ship.

PerplexityClaudeNotion
⧉ Brief
Goal

Competitor URLs → 3-tier pricing with feature diff and your positioning, ready to ship.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Competitor URLs → 3-tier pricing with feature diff and your positioning, ready to ship.
Starter prompt
Build "Pricing page generator" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Pricing page generator · 3 tools
51

Customer interview summarizer

Intermediate

5 interview transcripts → themes, quotes, and a Notion page the team can actually use.

WhisperClaudeNotion
⧉ Brief
Goal

5 interview transcripts → themes, quotes, and a Notion page the team can actually use.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →5 interview transcripts → themes, quotes, and a Notion page the team can actually use.
Starter prompt
Build "Customer interview summarizer" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Customer interview summarizer · 3 tools
52

A/B test copy generator

Intermediate

Page URL → 5 variants of the headline, subhead, and CTA, each with a hypothesis.

ClaudeMastraNotion
⧉ Brief
Goal

Page URL → 5 variants of the headline, subhead, and CTA, each with a hypothesis.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Page URL → 5 variants of the headline, subhead, and CTA, each with a hypothesis.
Starter prompt
Build "A/B test copy generator" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
A/B test copy generator · 3 tools
53

SEO audit agent

Advanced

Site URL → full audit: technical, on-page, backlinks, and a prioritized fix list.

ClaudeTavilyMastra
⧉ Brief
Goal

Site URL → full audit: technical, on-page, backlinks, and a prioritized fix list.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Site URL → full audit: technical, on-page, backlinks, and a prioritized fix list.
Starter prompt
Build "SEO audit agent" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
SEO audit agent · 3 tools
54

Testimonial collector

Intermediate

Review sites + G2 + your inbox → curated testimonials tagged by use case, ready for the website.

PerplexityClaudeNotion
⧉ Brief
Goal

Review sites + G2 + your inbox → curated testimonials tagged by use case, ready for the website.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Review sites + G2 + your inbox → curated testimonials tagged by use case, ready for the website.
Starter prompt
Build "Testimonial collector" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Testimonial collector · 3 tools
55

Sales call analyzer

Intermediate

Recording → transcript, top objections, winning phrases, and a coaching note for the rep.

WhisperClaudeNotion
⧉ Brief
Goal

Recording → transcript, top objections, winning phrases, and a coaching note for the rep.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Recording → transcript, top objections, winning phrases, and a coaching note for the rep.
Starter prompt
Build "Sales call analyzer" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Sales call analyzer · 3 tools
56

Proposal generator

Intermediate

Client brief + past wins → full proposal with scope, timeline, pricing, and case studies.

Claude Sonnet 5NotionGoogle Docs
⧉ Brief
Goal

Client brief + past wins → full proposal with scope, timeline, pricing, and case studies.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Client brief + past wins → full proposal with scope, timeline, pricing, and case studies.
Starter prompt
Build "Proposal generator" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Proposal generator · 3 tools
57

Churn predictor

Advanced

Usage events → at-risk user list with reasons and the offer most likely to retain them.

OpenClawCustom MLClaude
⧉ Brief
Goal

Usage events → at-risk user list with reasons and the offer most likely to retain them.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Usage events → at-risk user list with reasons and the offer most likely to retain them.
Starter prompt
Build "Churn predictor" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Churn predictor · 3 tools
58

Product launch checklist

Beginner

Launch date → 30-day checklist with copy, design, dev, and PR tasks, owners included.

ClaudeMastraNotion
⧉ Brief
Goal

Launch date → 30-day checklist with copy, design, dev, and PR tasks, owners included.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Launch date → 30-day checklist with copy, design, dev, and PR tasks, owners included.
Starter prompt
Build "Product launch checklist" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Product launch checklist · 3 tools
59

Brand voice analyzer

Advanced

10 pieces of your best content → voice profile: tone, sentence length, vocabulary, do/don't list.

ClaudeMastraNotion
⧉ Brief
Goal

10 pieces of your best content → voice profile: tone, sentence length, vocabulary, do/don't list.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →10 pieces of your best content → voice profile: tone, sentence length, vocabulary, do/don't list.
Starter prompt
Build "Brand voice analyzer" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Brand voice analyzer · 3 tools
60

Influencer finder

Intermediate

Niche + audience size + budget → list of 20 creators with overlap score and contact email.

PerplexityClaudeSheets
⧉ Brief
Goal

Niche + audience size + budget → list of 20 creators with overlap score and contact email.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Niche + audience size + budget → list of 20 creators with overlap score and contact email.
Starter prompt
Build "Influencer finder" as a daily-runner.
Input: <TRIGGER_EVENT>
Action: <WHAT_THE_AGENT_DOES>
Output: <WHERE_THE_RESULT_LIVES> (CRM, Notion, Sheets, Slack, etc.)
Constraint: anything below threshold <X> goes to a "review" queue, not the main flow.
Log every run to a dated file.
Influencer finder · 3 tools
Category 0510 systems

WhatsApp & Instagram

Where your customers already are. Reply before they tap away.

61

WhatsApp auto-replier

Intermediate

FAQ + business hours → instant reply, hands off to a human for anything outside the doc.

OpenClawWhatsApp APIClaude
⧉ Brief
Goal

FAQ + business hours → instant reply, hands off to a human for anything outside the doc.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →FAQ + business hours → instant reply, hands off to a human for anything outside the doc.
Starter prompt
Build "WhatsApp auto-replier".
Watch: <CHANNEL> (WhatsApp, IG DM, IG comment, etc.).
When: <EVENT>.
Reply with: <BEHAVIOUR>.
Hand off to a human when: <TRIGGER> (e.g., message is negative, request is outside FAQ, etc.).
Always include the original message in the human handoff.
WhatsApp auto-replier · 3 tools
62

Instagram DM qualifier

Intermediate

DM arrives → Claude scores intent, asks qualifying questions, books a call if hot.

ClaudeOpenClawMeta Graph API
⧉ Brief
Goal

DM arrives → Claude scores intent, asks qualifying questions, books a call if hot.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →DM arrives → Claude scores intent, asks qualifying questions, books a call if hot.
Starter prompt
Build "Instagram DM qualifier".
Watch: <CHANNEL> (WhatsApp, IG DM, IG comment, etc.).
When: <EVENT>.
Reply with: <BEHAVIOUR>.
Hand off to a human when: <TRIGGER> (e.g., message is negative, request is outside FAQ, etc.).
Always include the original message in the human handoff.
Instagram DM qualifier · 3 tools
63

Story poller

Advanced

Watches your story replies, drafts a personalized reply, sends if it fits the brand voice.

ClaudeOpenClawMeta Graph API
⧉ Brief
Goal

Watches your story replies, drafts a personalized reply, sends if it fits the brand voice.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Watches your story replies, drafts a personalized reply, sends if it fits the brand voice.
Starter prompt
Build "Story poller".
Watch: <CHANNEL> (WhatsApp, IG DM, IG comment, etc.).
When: <EVENT>.
Reply with: <BEHAVIOUR>.
Hand off to a human when: <TRIGGER> (e.g., message is negative, request is outside FAQ, etc.).
Always include the original message in the human handoff.
Story poller · 3 tools
64

Comment auto-responder

Advanced

New comment → Claude drafts a reply in your tone, flags anything negative for human review.

Claude Sonnet 5MastraMeta Graph API
⧉ Brief
Goal

New comment → Claude drafts a reply in your tone, flags anything negative for human review.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →New comment → Claude drafts a reply in your tone, flags anything negative for human review.
Starter prompt
Build "Comment auto-responder".
Watch: <CHANNEL> (WhatsApp, IG DM, IG comment, etc.).
When: <EVENT>.
Reply with: <BEHAVIOUR>.
Hand off to a human when: <TRIGGER> (e.g., message is negative, request is outside FAQ, etc.).
Always include the original message in the human handoff.
Comment auto-responder · 3 tools
65

Content cross-poster

Intermediate

One IG post → auto-formatted for LinkedIn, Twitter, Facebook, Threads, and Pinterest.

OpenClawMakeClaude
⧉ Brief
Goal

One IG post → auto-formatted for LinkedIn, Twitter, Facebook, Threads, and Pinterest.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →One IG post → auto-formatted for LinkedIn, Twitter, Facebook, Threads, and Pinterest.
Starter prompt
Build "Content cross-poster".
Watch: <CHANNEL> (WhatsApp, IG DM, IG comment, etc.).
When: <EVENT>.
Reply with: <BEHAVIOUR>.
Hand off to a human when: <TRIGGER> (e.g., message is negative, request is outside FAQ, etc.).
Always include the original message in the human handoff.
Content cross-poster · 3 tools
66

Influencer outreach

Intermediate

Creator handle → personalized DM pitch with a hook, your offer, and a one-tap yes.

ClaudeOpenClawMeta Graph API
⧉ Brief
Goal

Creator handle → personalized DM pitch with a hook, your offer, and a one-tap yes.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Creator handle → personalized DM pitch with a hook, your offer, and a one-tap yes.
Starter prompt
Build "Influencer outreach".
Watch: <CHANNEL> (WhatsApp, IG DM, IG comment, etc.).
When: <EVENT>.
Reply with: <BEHAVIOUR>.
Hand off to a human when: <TRIGGER> (e.g., message is negative, request is outside FAQ, etc.).
Always include the original message in the human handoff.
Influencer outreach · 3 tools
67

Hashtag research

Beginner

Topic + platform → 30 hashtags grouped by size, with predicted reach and competition.

PerplexityClaudeSheets
⧉ Brief
Goal

Topic + platform → 30 hashtags grouped by size, with predicted reach and competition.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Topic + platform → 30 hashtags grouped by size, with predicted reach and competition.
Starter prompt
Build "Hashtag research".
Watch: <CHANNEL> (WhatsApp, IG DM, IG comment, etc.).
When: <EVENT>.
Reply with: <BEHAVIOUR>.
Hand off to a human when: <TRIGGER> (e.g., message is negative, request is outside FAQ, etc.).
Always include the original message in the human handoff.
Hashtag research · 3 tools
68

Reel hook generator

Beginner

Topic + audience → 5 first-3-second hooks, each with a different emotional angle.

Claude Sonnet 5OpenClaw
⧉ Brief
Goal

Topic + audience → 5 first-3-second hooks, each with a different emotional angle.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Topic + audience → 5 first-3-second hooks, each with a different emotional angle.
Starter prompt
Build "Reel hook generator".
Watch: <CHANNEL> (WhatsApp, IG DM, IG comment, etc.).
When: <EVENT>.
Reply with: <BEHAVIOUR>.
Hand off to a human when: <TRIGGER> (e.g., message is negative, request is outside FAQ, etc.).
Always include the original message in the human handoff.
Reel hook generator · 2 tools
69

Caption A/B tester

Intermediate

Photo + goal → 5 caption variants, each with a different hook angle, ready to test.

ClaudeMastraNotion
⧉ Brief
Goal

Photo + goal → 5 caption variants, each with a different hook angle, ready to test.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Photo + goal → 5 caption variants, each with a different hook angle, ready to test.
Starter prompt
Build "Caption A/B tester".
Watch: <CHANNEL> (WhatsApp, IG DM, IG comment, etc.).
When: <EVENT>.
Reply with: <BEHAVIOUR>.
Hand off to a human when: <TRIGGER> (e.g., message is negative, request is outside FAQ, etc.).
Always include the original message in the human handoff.
Caption A/B tester · 3 tools
70

Story-to-post

Advanced

IG story screenshot → cleaned, reformatted, captioned post for the same audience.

Nano Banana 2ClaudeOpenClaw
⧉ Brief
Goal

IG story screenshot → cleaned, reformatted, captioned post for the same audience.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →IG story screenshot → cleaned, reformatted, captioned post for the same audience.
Starter prompt
Build "Story-to-post".
Watch: <CHANNEL> (WhatsApp, IG DM, IG comment, etc.).
When: <EVENT>.
Reply with: <BEHAVIOUR>.
Hand off to a human when: <TRIGGER> (e.g., message is negative, request is outside FAQ, etc.).
Always include the original message in the human handoff.
Story-to-post · 3 tools
Category 0610 systems

E-commerce & D2C

Shopify, D2C, marketplace. The agent does the boring work.

71

Product description generator

Intermediate

SKU + features → SEO description, bullet points, and meta in your brand voice.

ClaudeOpenClawShopify API
⧉ Brief
Goal

SKU + features → SEO description, bullet points, and meta in your brand voice.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →SKU + features → SEO description, bullet points, and meta in your brand voice.
Starter prompt
Build "Product description generator".
Source: <WHERE_THE_DATA_LIVES> (Shopify, Sheets, etc.).
Per item: <WHAT_TO_GENERATE_OR_DO>.
Output to: <DESTINATION>.
Never overwrite — write to a new column or a dated sheet.
Include the original SKU/ID in the output.
Product description generator · 3 tools
72

Product photo enhancement

Beginner

Casual phone shot → studio-style image: clean background, soft shadow, correct light.

Nano Banana 2OpenClaw
⧉ Brief
Goal

Casual phone shot → studio-style image: clean background, soft shadow, correct light.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Casual phone shot → studio-style image: clean background, soft shadow, correct light.
Starter prompt
Build "Product photo enhancement".
Source: <WHERE_THE_DATA_LIVES> (Shopify, Sheets, etc.).
Per item: <WHAT_TO_GENERATE_OR_DO>.
Output to: <DESTINATION>.
Never overwrite — write to a new column or a dated sheet.
Include the original SKU/ID in the output.
Product photo enhancement · 2 tools
73

Customer review summarizer

Intermediate

200 reviews → themes, top 3 complaints, top 3 wins, and a quote wall for the site.

ClaudeOpenClawNotion
⧉ Brief
Goal

200 reviews → themes, top 3 complaints, top 3 wins, and a quote wall for the site.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →200 reviews → themes, top 3 complaints, top 3 wins, and a quote wall for the site.
Starter prompt
Build "Customer review summarizer".
Source: <WHERE_THE_DATA_LIVES> (Shopify, Sheets, etc.).
Per item: <WHAT_TO_GENERATE_OR_DO>.
Output to: <DESTINATION>.
Never overwrite — write to a new column or a dated sheet.
Include the original SKU/ID in the output.
Customer review summarizer · 3 tools
74

Inventory alert

Intermediate

Stock crosses threshold → Telegram alert to the founder, with the SKU and reorder link.

OpenClawShopify APITelegram MCP
⧉ Brief
Goal

Stock crosses threshold → Telegram alert to the founder, with the SKU and reorder link.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Stock crosses threshold → Telegram alert to the founder, with the SKU and reorder link.
Starter prompt
Build "Inventory alert".
Source: <WHERE_THE_DATA_LIVES> (Shopify, Sheets, etc.).
Per item: <WHAT_TO_GENERATE_OR_DO>.
Output to: <DESTINATION>.
Never overwrite — write to a new column or a dated sheet.
Include the original SKU/ID in the output.
Inventory alert · 3 tools
75

Cart abandonment email

Intermediate

Abandoned cart → 3-email sequence: reminder, social proof, last-chance offer.

Claude Sonnet 5OpenClawKlaviyo
⧉ Brief
Goal

Abandoned cart → 3-email sequence: reminder, social proof, last-chance offer.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Abandoned cart → 3-email sequence: reminder, social proof, last-chance offer.
Starter prompt
Build "Cart abandonment email".
Source: <WHERE_THE_DATA_LIVES> (Shopify, Sheets, etc.).
Per item: <WHAT_TO_GENERATE_OR_DO>.
Output to: <DESTINATION>.
Never overwrite — write to a new column or a dated sheet.
Include the original SKU/ID in the output.
Cart abandonment email · 3 tools
76

Bundle recommender

Advanced

Items in cart → 2 add-on bundles that historically convert with this combo.

ClaudeCustom MLShopify API
⧉ Brief
Goal

Items in cart → 2 add-on bundles that historically convert with this combo.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Items in cart → 2 add-on bundles that historically convert with this combo.
Starter prompt
Build "Bundle recommender".
Source: <WHERE_THE_DATA_LIVES> (Shopify, Sheets, etc.).
Per item: <WHAT_TO_GENERATE_OR_DO>.
Output to: <DESTINATION>.
Never overwrite — write to a new column or a dated sheet.
Include the original SKU/ID in the output.
Bundle recommender · 3 tools
77

Returns processor

Advanced

Return reason text → accept/decline decision + suggested action + customer reply.

ClaudeOpenClawShopify API
⧉ Brief
Goal

Return reason text → accept/decline decision + suggested action + customer reply.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Return reason text → accept/decline decision + suggested action + customer reply.
Starter prompt
Build "Returns processor".
Source: <WHERE_THE_DATA_LIVES> (Shopify, Sheets, etc.).
Per item: <WHAT_TO_GENERATE_OR_DO>.
Output to: <DESTINATION>.
Never overwrite — write to a new column or a dated sheet.
Include the original SKU/ID in the output.
Returns processor · 3 tools
78

Loyalty program manager

Intermediate

Purchase → tier update, points credit, and a personalized nudge toward the next reward.

OpenClawCustom logicSheets
⧉ Brief
Goal

Purchase → tier update, points credit, and a personalized nudge toward the next reward.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Purchase → tier update, points credit, and a personalized nudge toward the next reward.
Starter prompt
Build "Loyalty program manager".
Source: <WHERE_THE_DATA_LIVES> (Shopify, Sheets, etc.).
Per item: <WHAT_TO_GENERATE_OR_DO>.
Output to: <DESTINATION>.
Never overwrite — write to a new column or a dated sheet.
Include the original SKU/ID in the output.
Loyalty program manager · 3 tools
79

Price competitor monitor

Advanced

Watch list of competitor URLs → Slack alert when their price drops or stock returns.

OpenClawTavilySheets
⧉ Brief
Goal

Watch list of competitor URLs → Slack alert when their price drops or stock returns.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Watch list of competitor URLs → Slack alert when their price drops or stock returns.
Starter prompt
Build "Price competitor monitor".
Source: <WHERE_THE_DATA_LIVES> (Shopify, Sheets, etc.).
Per item: <WHAT_TO_GENERATE_OR_DO>.
Output to: <DESTINATION>.
Never overwrite — write to a new column or a dated sheet.
Include the original SKU/ID in the output.
Price competitor monitor · 3 tools
80

SKU image generator

Intermediate

Single product photo → 4 hero angles, lifestyle shot, and detail close-up.

Nano Banana 2ClaudeShopify API
⧉ Brief
Goal

Single product photo → 4 hero angles, lifestyle shot, and detail close-up.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Single product photo → 4 hero angles, lifestyle shot, and detail close-up.
Starter prompt
Build "SKU image generator".
Source: <WHERE_THE_DATA_LIVES> (Shopify, Sheets, etc.).
Per item: <WHAT_TO_GENERATE_OR_DO>.
Output to: <DESTINATION>.
Never overwrite — write to a new column or a dated sheet.
Include the original SKU/ID in the output.
SKU image generator · 3 tools
Category 0710 systems

Research & Learning

Read 10x faster. Learn what matters. Skip the rest.

81

Paper summarizer

Beginner

ArXiv PDF or DOI → 1-page summary: problem, method, results, limitations, 3 questions.

NotebookLMClaudeTavily
⧉ Brief
Goal

ArXiv PDF or DOI → 1-page summary: problem, method, results, limitations, 3 questions.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →ArXiv PDF or DOI → 1-page summary: problem, method, results, limitations, 3 questions.
Starter prompt
Build "Paper summarizer".
Source: <URL / PDF / DOC>.
Length: <WORDS>.
Format: <BULLETS / PARAGRAPHS / TABLE>.
Cite every claim. If you can't cite it, say "I don't know" — don't make it up.
End with 3 questions the user should ask next.
Paper summarizer · 3 tools
82

Topic deep-dive

Intermediate

Topic → 10-source research report with citations, contradictions, and a reading order.

PerplexityClaudeNotion
⧉ Brief
Goal

Topic → 10-source research report with citations, contradictions, and a reading order.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Topic → 10-source research report with citations, contradictions, and a reading order.
Starter prompt
Build "Topic deep-dive".
Source: <URL / PDF / DOC>.
Length: <WORDS>.
Format: <BULLETS / PARAGRAPHS / TABLE>.
Cite every claim. If you can't cite it, say "I don't know" — don't make it up.
End with 3 questions the user should ask next.
Topic deep-dive · 3 tools
83

Book notes generator

Beginner

Chapter PDF or text → structured notes: key ideas, quotes, action steps, and links.

NotebookLMClaudeNotion
⧉ Brief
Goal

Chapter PDF or text → structured notes: key ideas, quotes, action steps, and links.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Chapter PDF or text → structured notes: key ideas, quotes, action steps, and links.
Starter prompt
Build "Book notes generator".
Source: <URL / PDF / DOC>.
Length: <WORDS>.
Format: <BULLETS / PARAGRAPHS / TABLE>.
Cite every claim. If you can't cite it, say "I don't know" — don't make it up.
End with 3 questions the user should ask next.
Book notes generator · 3 tools
84

Concept explainer

Beginner

Jargon term + your level → plain explanation, an analogy, and a worked example.

ClaudeNotebookLM
⧉ Brief
Goal

Jargon term + your level → plain explanation, an analogy, and a worked example.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Jargon term + your level → plain explanation, an analogy, and a worked example.
Starter prompt
Build "Concept explainer".
Source: <URL / PDF / DOC>.
Length: <WORDS>.
Format: <BULLETS / PARAGRAPHS / TABLE>.
Cite every claim. If you can't cite it, say "I don't know" — don't make it up.
End with 3 questions the user should ask next.
Concept explainer · 2 tools
85

Quiz generator

Intermediate

Paste study material → 20 multiple-choice questions with answers and one-line explanations.

ClaudeMastraNotion
⧉ Brief
Goal

Paste study material → 20 multiple-choice questions with answers and one-line explanations.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Paste study material → 20 multiple-choice questions with answers and one-line explanations.
Starter prompt
Build "Quiz generator".
Source: <URL / PDF / DOC>.
Length: <WORDS>.
Format: <BULLETS / PARAGRAPHS / TABLE>.
Cite every claim. If you can't cite it, say "I don't know" — don't make it up.
End with 3 questions the user should ask next.
Quiz generator · 3 tools
86

Citation finder

Intermediate

Claim + field → 5 peer-reviewed sources that back it up, with a confidence note on each.

PerplexityClaudeTavily
⧉ Brief
Goal

Claim + field → 5 peer-reviewed sources that back it up, with a confidence note on each.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Claim + field → 5 peer-reviewed sources that back it up, with a confidence note on each.
Starter prompt
Build "Citation finder".
Source: <URL / PDF / DOC>.
Length: <WORDS>.
Format: <BULLETS / PARAGRAPHS / TABLE>.
Cite every claim. If you can't cite it, say "I don't know" — don't make it up.
End with 3 questions the user should ask next.
Citation finder · 3 tools
87

Reading list curator

Intermediate

Goal + level → 10 books/papers/courses in priority order, with a 'why this one' per item.

ClaudeTavilyNotion
⧉ Brief
Goal

Goal + level → 10 books/papers/courses in priority order, with a 'why this one' per item.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Goal + level → 10 books/papers/courses in priority order, with a 'why this one' per item.
Starter prompt
Build "Reading list curator".
Source: <URL / PDF / DOC>.
Length: <WORDS>.
Format: <BULLETS / PARAGRAPHS / TABLE>.
Cite every claim. If you can't cite it, say "I don't know" — don't make it up.
End with 3 questions the user should ask next.
Reading list curator · 3 tools
88

Language tutor

Intermediate

Spanish learner → 10-minute practice: read, listen (ElevenLabs), respond, get feedback.

ClaudeElevenLabsMake
⧉ Brief
Goal

Spanish learner → 10-minute practice: read, listen (ElevenLabs), respond, get feedback.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Spanish learner → 10-minute practice: read, listen (ElevenLabs), respond, get feedback.
Starter prompt
Build "Language tutor".
Source: <URL / PDF / DOC>.
Length: <WORDS>.
Format: <BULLETS / PARAGRAPHS / TABLE>.
Cite every claim. If you can't cite it, say "I don't know" — don't make it up.
End with 3 questions the user should ask next.
Language tutor · 3 tools
89

Speed reader

Beginner

30-page doc → 1-page TL;DR with the 3 must-knows, the 5 supporting points, and the 1 caveat.

ClaudeMastraNotebookLM
⧉ Brief
Goal

30-page doc → 1-page TL;DR with the 3 must-knows, the 5 supporting points, and the 1 caveat.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →30-page doc → 1-page TL;DR with the 3 must-knows, the 5 supporting points, and the 1 caveat.
Starter prompt
Build "Speed reader".
Source: <URL / PDF / DOC>.
Length: <WORDS>.
Format: <BULLETS / PARAGRAPHS / TABLE>.
Cite every claim. If you can't cite it, say "I don't know" — don't make it up.
End with 3 questions the user should ask next.
Speed reader · 3 tools
90

Concept map generator

Advanced

Topic → visual mind map image with branches, sub-topics, and a one-line description per node.

ClaudeNano Banana 2Notion
⧉ Brief
Goal

Topic → visual mind map image with branches, sub-topics, and a one-line description per node.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Topic → visual mind map image with branches, sub-topics, and a one-line description per node.
Starter prompt
Build "Concept map generator".
Source: <URL / PDF / DOC>.
Length: <WORDS>.
Format: <BULLETS / PARAGRAPHS / TABLE>.
Cite every claim. If you can't cite it, say "I don't know" — don't make it up.
End with 3 questions the user should ask next.
Concept map generator · 3 tools
Category 085 systems

Health & Lifestyle

Habits, meals, sleep. Boring stuff, automated so you actually do it.

91

Workout plan generator

Beginner

Goal + equipment + 30 min/day → 4-week plan with progression and rest day logic.

ClaudeOpenClawNotion
⧉ Brief
Goal

Goal + equipment + 30 min/day → 4-week plan with progression and rest day logic.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Goal + equipment + 30 min/day → 4-week plan with progression and rest day logic.
Starter prompt
Build "Workout plan generator" as a personal coach.
Profile: <AGE, GOAL, EQUIPMENT, TIME_PER_DAY>.
Generate: <WEEKLY_PLAN, DAILY_REMINDER, etc.>.
Rules:
- Rest days are non-negotiable.
- Progress by 5%/week, not by 50%.
- Skip the previous day if I miss it — never skip ahead.
Workout plan generator · 3 tools
92

Meal planner

Intermediate

Diet + budget + dislikes → 7-day meal plan with grocery list and prep instructions.

ClaudeOpenClawSheets
⧉ Brief
Goal

Diet + budget + dislikes → 7-day meal plan with grocery list and prep instructions.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Diet + budget + dislikes → 7-day meal plan with grocery list and prep instructions.
Starter prompt
Build "Meal planner" as a personal coach.
Profile: <AGE, GOAL, EQUIPMENT, TIME_PER_DAY>.
Generate: <WEEKLY_PLAN, DAILY_REMINDER, etc.>.
Rules:
- Rest days are non-negotiable.
- Progress by 5%/week, not by 50%.
- Skip the previous day if I miss it — never skip ahead.
Meal planner · 3 tools
93

Sleep tracker

Advanced

Sleep data → insights on patterns, a recommended bedtime, and a wind-down routine.

OpenClawCustom logicClaude
⧉ Brief
Goal

Sleep data → insights on patterns, a recommended bedtime, and a wind-down routine.

Time

~1–2 days · Advanced

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Sleep data → insights on patterns, a recommended bedtime, and a wind-down routine.
Starter prompt
Build "Sleep tracker" as a personal coach.
Profile: <AGE, GOAL, EQUIPMENT, TIME_PER_DAY>.
Generate: <WEEKLY_PLAN, DAILY_REMINDER, etc.>.
Rules:
- Rest days are non-negotiable.
- Progress by 5%/week, not by 50%.
- Skip the previous day if I miss it — never skip ahead.
Sleep tracker · 3 tools
94

Mood journal companion

Beginner

Daily mood + one line → weekly insights: triggers, what helped, one experiment to try.

ClaudeOpenClawNotion
⧉ Brief
Goal

Daily mood + one line → weekly insights: triggers, what helped, one experiment to try.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Daily mood + one line → weekly insights: triggers, what helped, one experiment to try.
Starter prompt
Build "Mood journal companion" as a personal coach.
Profile: <AGE, GOAL, EQUIPMENT, TIME_PER_DAY>.
Generate: <WEEKLY_PLAN, DAILY_REMINDER, etc.>.
Rules:
- Rest days are non-negotiable.
- Progress by 5%/week, not by 50%.
- Skip the previous day if I miss it — never skip ahead.
Mood journal companion · 3 tools
95

Habit streak coach

Beginner

Streak count + last slip → motivation message, a small reset plan, and a one-tap nudge.

ClaudeOpenClawTelegram MCP
⧉ Brief
Goal

Streak count + last slip → motivation message, a small reset plan, and a one-tap nudge.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Streak count + last slip → motivation message, a small reset plan, and a one-tap nudge.
Starter prompt
Build "Habit streak coach" as a personal coach.
Profile: <AGE, GOAL, EQUIPMENT, TIME_PER_DAY>.
Generate: <WEEKLY_PLAN, DAILY_REMINDER, etc.>.
Rules:
- Rest days are non-negotiable.
- Progress by 5%/week, not by 50%.
- Skip the previous day if I miss it — never skip ahead.
Habit streak coach · 3 tools
Category 095 systems

Real Estate & Local Business

Local shops, brokers, clinics. The agent does the follow-up.

96

Listing description generator

Beginner

Property details + photos → listing copy with headline, body, neighbourhood, and CTA.

ClaudeNano Banana 2Notion
⧉ Brief
Goal

Property details + photos → listing copy with headline, body, neighbourhood, and CTA.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Property details + photos → listing copy with headline, body, neighbourhood, and CTA.
Starter prompt
Build "Listing description generator".
Inputs: <PROPERTY / LEAD / FAQ>.
Output: <LISTING / WHATSAPP_SEQ / BOOKING_LINK>.
Tone: <LOCAL_TRUST, NOT_CORPORATE>.
Always include the agent's phone number at the bottom.
If a question is outside the FAQ, escalate to a human, don't guess.
Listing description generator · 3 tools
97

Lead follow-up

Intermediate

Enquiry → 5-touch WhatsApp sequence over 7 days, branched by reply behaviour.

ClaudeOpenClawWhatsApp API
⧉ Brief
Goal

Enquiry → 5-touch WhatsApp sequence over 7 days, branched by reply behaviour.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Enquiry → 5-touch WhatsApp sequence over 7 days, branched by reply behaviour.
Starter prompt
Build "Lead follow-up".
Inputs: <PROPERTY / LEAD / FAQ>.
Output: <LISTING / WHATSAPP_SEQ / BOOKING_LINK>.
Tone: <LOCAL_TRUST, NOT_CORPORATE>.
Always include the agent's phone number at the bottom.
If a question is outside the FAQ, escalate to a human, don't guess.
Lead follow-up · 3 tools
98

Tenant Q&A bot

Intermediate

FAQ + your lease terms → instant answers, escalates only on edge cases.

ClaudeOpenClawWhatsApp API
⧉ Brief
Goal

FAQ + your lease terms → instant answers, escalates only on edge cases.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →FAQ + your lease terms → instant answers, escalates only on edge cases.
Starter prompt
Build "Tenant Q&A bot".
Inputs: <PROPERTY / LEAD / FAQ>.
Output: <LISTING / WHATSAPP_SEQ / BOOKING_LINK>.
Tone: <LOCAL_TRUST, NOT_CORPORATE>.
Always include the agent's phone number at the bottom.
If a question is outside the FAQ, escalate to a human, don't guess.
Tenant Q&A bot · 3 tools
99

Appointment scheduler

Intermediate

Free slots in Calendar → booking link, confirmation, reminder, and a reschedule path.

ClaudeCalendar MCPMake
⧉ Brief
Goal

Free slots in Calendar → booking link, confirmation, reminder, and a reschedule path.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Free slots in Calendar → booking link, confirmation, reminder, and a reschedule path.
Starter prompt
Build "Appointment scheduler".
Inputs: <PROPERTY / LEAD / FAQ>.
Output: <LISTING / WHATSAPP_SEQ / BOOKING_LINK>.
Tone: <LOCAL_TRUST, NOT_CORPORATE>.
Always include the agent's phone number at the bottom.
If a question is outside the FAQ, escalate to a human, don't guess.
Appointment scheduler · 3 tools
100

Local SEO content

Intermediate

City + service → 1,500-word blog post optimized for 'near me' and local keywords.

ClaudePerplexityWordPress
⧉ Brief
Goal

City + service → 1,500-word blog post optimized for 'near me' and local keywords.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →City + service → 1,500-word blog post optimized for 'near me' and local keywords.
Starter prompt
Build "Local SEO content".
Inputs: <PROPERTY / LEAD / FAQ>.
Output: <LISTING / WHATSAPP_SEQ / BOOKING_LINK>.
Tone: <LOCAL_TRUST, NOT_CORPORATE>.
Always include the agent's phone number at the bottom.
If a question is outside the FAQ, escalate to a human, don't guess.
Local SEO content · 3 tools
Category 105 systems

Education & Students

Study smarter, not longer. For students and lifelong learners.

101

Study plan generator

Beginner

Subject + exam date + weak areas → 4-week study plan with daily tasks and checkpoints.

ClaudeNotebookLMNotion
⧉ Brief
Goal

Subject + exam date + weak areas → 4-week study plan with daily tasks and checkpoints.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Subject + exam date + weak areas → 4-week study plan with daily tasks and checkpoints.
Starter prompt
Build "Study plan generator" for a student.
Subject: <TOPIC>.
Level: <GRADE_YEAR>.
Format: <STUDY_PLAN / FLASHCARDS / OUTLINE>.
Keep it short — the student has 30 minutes a day, not 3 hours.
Include 1 practice question at the end of every section.
Study plan generator · 3 tools
102

Flashcard generator

Beginner

Notes or chapter → 30 flashcards with question, answer, and a spaced-repetition schedule.

ClaudeMastraAnki
⧉ Brief
Goal

Notes or chapter → 30 flashcards with question, answer, and a spaced-repetition schedule.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Notes or chapter → 30 flashcards with question, answer, and a spaced-repetition schedule.
Starter prompt
Build "Flashcard generator" for a student.
Subject: <TOPIC>.
Level: <GRADE_YEAR>.
Format: <STUDY_PLAN / FLASHCARDS / OUTLINE>.
Keep it short — the student has 30 minutes a day, not 3 hours.
Include 1 practice question at the end of every section.
Flashcard generator · 3 tools
103

Essay outline

Beginner

Topic + word count + level → 5-paragraph outline with thesis, evidence, and counter-arguments.

Claude Sonnet 5Notion
⧉ Brief
Goal

Topic + word count + level → 5-paragraph outline with thesis, evidence, and counter-arguments.

Time

~2 hours · Beginner

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Topic + word count + level → 5-paragraph outline with thesis, evidence, and counter-arguments.
Starter prompt
Build "Essay outline" for a student.
Subject: <TOPIC>.
Level: <GRADE_YEAR>.
Format: <STUDY_PLAN / FLASHCARDS / OUTLINE>.
Keep it short — the student has 30 minutes a day, not 3 hours.
Include 1 practice question at the end of every section.
Essay outline · 2 tools
104

Lab report formatter

Intermediate

Raw data + hypothesis → formatted lab report with sections, tables, and a discussion draft.

ClaudeNano Banana 2Google Docs
⧉ Brief
Goal

Raw data + hypothesis → formatted lab report with sections, tables, and a discussion draft.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Raw data + hypothesis → formatted lab report with sections, tables, and a discussion draft.
Starter prompt
Build "Lab report formatter" for a student.
Subject: <TOPIC>.
Level: <GRADE_YEAR>.
Format: <STUDY_PLAN / FLASHCARDS / OUTLINE>.
Keep it short — the student has 30 minutes a day, not 3 hours.
Include 1 practice question at the end of every section.
Lab report formatter · 3 tools
105

Math tutor

Intermediate

Problem → step-by-step solution in plain language, with a similar problem to try next.

Claude Sonnet 5Notion
⧉ Brief
Goal

Problem → step-by-step solution in plain language, with a similar problem to try next.

Time

~half a day · Intermediate

Inputs
  • →(whatever the description mentions as the trigger)
Outputs
  • →Problem → step-by-step solution in plain language, with a similar problem to try next.
Starter prompt
Build "Math tutor" for a student.
Subject: <TOPIC>.
Level: <GRADE_YEAR>.
Format: <STUDY_PLAN / FLASHCARDS / OUTLINE>.
Keep it short — the student has 30 minutes a day, not 3 hours.
Include 1 practice question at the end of every section.
Math tutor · 2 tools
Methodology

How to build any of these.

Same 3 steps. Works for every system on this page. No matter your level.
Step 01

Pick a system

Browse the 100+ below. Pick one that solves a real problem you have this week. Not a cool demo — a real annoyance.

Step 02

Get the tools (all free or open)

Every system on this page uses free or open-source tools. Claude Sonnet 5 (free tier), GPT-5.6 (free tier), OpenClaw (self-hosted, 210K+ stars), n8n (self-hosted), Nano Banana 2 (free tier), ElevenLabs (free tier). Total stack cost: ₹0.

Step 03

Build with Claude Code + OpenClaw

Use Claude Code as the brain, OpenClaw as the wiring. 80% of the systems on this page are a Claude prompt + a 4-node OpenClaw flow. Ship in a weekend.

Want us to build it for you?

We build custom agentic systems for Indian SMBs. 2-week sprints. Fixed scope. We ship the system and train your team to run it.

Get the Agents Setup service →

Want a team workshop?

Bring 5-20 people. We run a 1-3 day custom workshop. Your team walks away with working systems they built themselves.

Book a custom workshop →
The 5 patterns

Every agent follows one of these 5 patterns.

When you strip away the marketing, every agentic system on this page uses one of these loops. Learn the 5, and you can build any of the 100+ above.
Pattern 01

Sense → Think → Act

The ReAct loop

The agent reads the world, reasons about the next step, takes one action, repeats. Used for: customer support, lead qualifiers, FAQ bots. Most beginner systems follow this pattern.

loop {
  observation = read_world()       // Sense
  thought     = llm(observation)   // Think
  action      = parse(thought)     // Decide
  apply(action)                    // Act
  if done(action): break
}
Pattern 02

Plan → Execute → Reflect

Plan-and-Execute

The agent drafts a full plan first, executes it step by step, then critiques its own output. Used for: research, content repurposing, code refactors.

plan = llm("draft a 5-step plan for: " + goal)
for step in plan:
  result = execute(step)
  log(result)
critique = llm("what's missing or wrong?")
if critique.needs_fix:
  replan(critique)
Pattern 03

Multi-agent role assignment

CrewAI pattern

Three or four specialized agents, each with a role (researcher, writer, editor, critic), pass work to each other. Used for: blog + image bundle, podcast show notes, social campaigns.

researcher = Agent("researcher", llm, tools=[search])
writer     = Agent("writer",     llm, style="brand voice")
editor     = Agent("editor",     llm, style="concise")
critic     = Agent("critic",     llm, style="tough")

brief   = researcher.run(topic)
draft   = writer.run(brief)
clean   = editor.run(draft)
ship    = critic.run(clean)
Pattern 04

Memory + Tools + Plan

LangGraph pattern

Stateful graph with persistent memory, tool access, and explicit plan nodes. Used for: long-running workflows, complex multi-step tasks, anything that needs to survive a restart.

graph = StateGraph({
  memory:  postgres,        // survives restarts
  tools:   [search, sql, mail],
  plan:    plan_node,
  execute: execute_node,
  reflect: reflect_node,
})
Pattern 05

Hooks + Skills + Loop

OpenClaw pattern

Trigger hooks fire on events (new email, new lead, new file), load the right skill, run the loop until done. Used for: WhatsApp auto-repliers, inventory alerts, content cross-posters.

on "new_email"   -> load("inbox-triage")  -> run(loop)
on "new_lead"    -> load("lead-qualifier") -> run(loop)
on "stock_low"   -> load("reorder-alert")  -> run(loop)
on "schedule 9am"-> load("daily-briefing") -> run(loop)
Pattern picker

Which pattern for which use case?

Match the system you want to build to the pattern that fits. The wrong pattern is the #1 reason beginners get stuck.
PatternUse when…FrameworksExample systems
Sense → Think → ActSingle goal, clear input, repeatable decisions.OpenClaw, LangGraph (basic), MastraWhatsApp auto-replier (system #61), Lead qualifier (#46), Inbox zero (#10)
Plan → Execute → ReflectMulti-step task, the agent needs to see the whole picture first.LangGraph, Mastra, Claude Agent SDKBlog post + image bundle (#37), Repurpose content (#38), Spec-to-code (#27)
Multi-agent role assignmentQuality matters more than speed. Each pass improves the output.CrewAI, AutoGen, LangGraph multi-agentPodcast show notes (#34), Customer interview summarizer (#51), Quote cards (#43)
Memory + Tools + PlanLong-running workflows, anything that needs to survive a restart.LangGraph, Mastra, Pydantic AIDaily standup bot (#15), Churn predictor (#57), SEO audit agent (#53)
Hooks + Skills + LoopEvent-driven, fires on triggers, runs unattended.OpenClaw, n8n, MakeInventory alert (#74), Daily morning briefing (#1), Price competitor monitor (#79)
AEO FAQ

Questions people actually ask about agentic systems.

Plain English answers. Real tool names. If a question isn’t here, mail us.

What is an agentic system?

+

An agentic system is software that uses an AI model (Claude, GPT-5.6, Gemini) to make decisions and take actions on its own — not just respond to prompts. It has tools, memory, and a goal. A chatbot answers a question; an agent reads your email, drafts a reply, files it in Notion, and pings you on Telegram. The 100+ systems on this page are all agentic — they do work, not just talk.

What are the best agentic systems for beginners?

+

Start with systems that have one input, one output, and three or fewer tools. Good first builds: daily AI morning briefing (Claude + Telegram MCP), meeting notes → action items (Whisper + GPT-5.6), recipe generator from fridge photo (Nano Banana 2 + Claude), tweet thread writer (GPT-5.6 + Claude). All four are Beginner difficulty on this page, take under 2 hours to build, and use only free tools.

How do I build an AI agent without coding?

+

Use OpenClaw (self-hosted, 210K+ GitHub stars) or n8n (also self-hosted) as the wiring. Use Claude Sonnet 5 (free tier) as the brain. The whole flow is drag-and-drop nodes plus a prompt. Most Beginner systems on this page (systems 1-8, 21, 29, 30, 35, 36, 40, 41, 43, 48, 58, 67, 68, 81, 83, 84, 89, 91, 94, 95, 96, 101, 102, 103) are buildable without writing code.

What's the best framework for building agents in 2026?

+

For self-hosted, privacy-first builds (Indian SMBs, schools, D2C brands): OpenClaw. For visual drag-and-drop: n8n or Make. For production agents with state and memory: LangGraph or Mastra. For quick prototypes: OpenClaw with Claude Sonnet 5. 2BFT's default stack is OpenClaw + Claude Sonnet 5 + Nano Banana 2 — it runs locally, costs nothing, and survives a power cut.

How many agentic systems can I build in a week?

+

A focused builder can ship 3-5 working systems in a week. In the 2BFT Academy's Vibe Coding Hackathon (Zoho Corporation, June 14-21 2026), builders shipped working AI agents, WhatsApp bots, and content pipelines in 8 days. Beginners should expect 1-2 per week. Pick a system, build it for a real use case, then add the next. Don't stack 10 unfinished demos.

What tools do I need to build agentic systems?

+

The minimum viable stack: one AI model (Claude Sonnet 5 free tier, GPT-5.6 free tier, or Gemini 3.1 Pro free tier), one orchestrator (OpenClaw self-hosted, n8n self-hosted, or Make free tier), and one source of truth (Notion, Google Sheets, or a Postgres database). Everything else is optional. Most systems on this page use 3-4 tools. The full free stack costs ₹0 to run.

How to use a brief

Click ⧉ Brief on any system. Copy. Paste. Build.

Each system above is a <details> block. Click the ⧉ Brief badge to expand the full brief: goal, inputs, outputs, tools, time estimate, and a starter prompt. Then hit Copy full brief and paste the whole thing into Claude Code, OpenClaw, or Cursor.

// pick your next move

100+ systems.
Pick one.

The 100+ above are not a brochure. They’re the systems 2BFT runs, teaches, and ships. Pick one, build it this week, and you’ll have a working agent by Friday.

Want us to build one for you? →Want a workshop for your team? →Want to learn from us? Mail us →

Built from Vaniyambadi, Tamil Nadu. ₹0 to start. 100+ to choose from.