Receipts
Changes
Business model · Agent-economy plays

Connectors & plugins for personal agents what founders report, what it costs, how to start

Connectors & plugins for personal agents (Meta Muse, OpenAI Dots, ChatGPT apps)

Expose a real-world service to personal agents (court bookings, repair dispatch, supplier leads, grocery lists) and earn per booking, lead or subscription.

Step 4 of 5 · Build an asset Open
Checked4 days ago13 OCT 2026 · every 60 days

Grade C: Self-reported. The grade says how well-sourced this page is, not how good Connectors & plugins for personal agents is. How well-sourced a page or fact is. Not a rating of the company. More in the glossary

Updated Checked Checked: we looked again and it was still right. Updated: something changed. More in the glossary

Personal-agent platforms opened developer programs in 2026 (Muse connectors via API or MCP with Meta review; OpenAI Dots plugins; 'Sign in with ChatGPT' lets users spend their own token allowance in your app, so a free core app costs you no inference). Ideas from the sources: lead gen for B2B suppliers (100 customers × $99/mo), appliance-repair dispatch (~$100 per booking), local padel court/match finder, family dinner → shoppable grocery list, 'real-world APIs' of vetted specialists (permit expediters, customs brokers) for a transaction fee, and analytics for agent discovery.

How Connectors & plugins for personal agents works

  1. Pick a service agents need mid-request (bookings, quotes, leads)
  2. Write one customer outcome; build quote and availability first
  3. Expose it via API or MCP; handle auth, idempotency and edge cases
  4. Submit to platform directories; distribute via creators and partners
  • DevDay launches (per podcast): Dots personal agent platform w/ plugins; Decisions API; Agents API w/ computer use; Sign in with ChatGPT using user's token allowance 2 Marketing claim · grade C (self-reported) · 13 Oct 2026

What founders report earning

Monthly potential $0–10,000/mo after ramp-up, part-time
Ceiling Very high

Pay figures are what platforms advertise or what workers report, each with its date. They are not typical or guaranteed earnings: yours depend on your skills, hours, location and how much work is available.

What it costs to start

Startup cost $0
Hours from 10 h a week async
  • Muse connector ideas & pricing: Supplier lead gen 100 × $99/mo; repair dispatch ~$100/booking; padel finder booking fee; Instacart shoppable lists 1 Marketing claim · grade D (unconfirmed) · 13 Oct 2026

How to start

  1. Test a manual sample (e.g., a small batch of leads) before building
  2. Test awkward cases (expired quote, double booking) before submitting

Is it still working in 2026?

Evidence: None No recent case we know of. How many founders reported success with this model in the last 12 months. More in the glossary

Cases in the last 12 months 0
Latest case none yet

AI outlook · emerging Distribution inside agent directories is unproven; plan your own channel (creators, product-led sharing, marketplaces).

What to watch

  • Platform review and policy risk
  • Economics are unclear while platforms evolve

Lawsuits and complaints are allegations unless a court has ruled. We report their status with dates and sources and update it when it changes.

Where people come from

The proof · #agent_connectors Grade C · checked 13 Oct 2026

The proof

Every fact on this page with its type, grade, date and source. Report an error

Sourced facts 2 facts · type · grade · as of

  • Marketing claim Muse connector ideas & pricing

    Supplier lead gen 100 × $99/mo; repair dispatch ~$100/booking; padel finder booking fee; Instacart shoppable lists 1

  • Marketing claim DevDay launches (per podcast)

    Dots personal agent platform w/ plugins; Decisions API; Agents API w/ computer use; Sign in with ChatGPT using user's token allowance 2

Sources 3 sources · type

  1. Meta Muse AI Connectors: The Next App Store Moment Community Greg Isenberg — Startup Ideas Podcast · podcast · accessed 13 Oct 2026 · published 2026 Private copy, not linked. Find on YouTube
  2. OpenAI DevDay: Dots, Agents & $100B Opportunities Community Greg Isenberg — Startup Ideas Podcast · podcast · accessed 13 Oct 2026 · published 2026 Private copy, not linked. Find on YouTube
  3. The Next $100B Market: Selling To AI Agents Community Greg Isenberg — Startup Ideas Podcast · podcast · accessed 13 Oct 2026 · published 2026 Private copy, not linked. Find on YouTube

Grade C (self-reported) · checked 13 Oct 2026 · re-checked every 60 days · how grades work · report an error

Changelog

Every change to this page, newest first. In the full changelog

  1. added all

    manual research (Claude) + sources listed · reviewer: owner review pending · grade C ·

Keep exploring

On the ladder

  1. 1 Paid tasks 25
  2. 2 Sell your skill 16
  3. 3 Run a service 12
  4. 4 Build an asset 20
  5. 5 Own a business 3

Step 4 of 5 · Build an asset. A product or audience that earns without your hours. Slow start, no ceiling.

More like this

Related

  • Step 3 · Service Open

    WebMCP / agent-conversion agency & 'agent mystery shopper'

    Make boring business websites usable by visitors' AI agents (WebMCP tools for quotes/bookings) and sell monthly 'can an agent complete the job?' test reports.

    Monthly potential $500–10,000/mo
    First money 30–120 days
    Cost to start $0
    Grade D means: Unconfirmed. The grade says how well-sourced this page is, not how good WebMCP / agent-conversion agency & 'agent mystery shopper' is. More in the glossary Unconfirmed Checked 4 days ago
  • Step 4 · Asset Open

    Niche data refinery ('clean fuel for agents')

    Track 100 businesses in one niche + city (prices, reviews, hiring, ads), turn it into reports, then a dashboard, API, MCP tool and eventually pay-per-lookup for agents.

    Monthly potential $0–15,000/mo
    First money 30–120 days
    Cost to start $0
    Grade C means: Self-reported. The grade says how well-sourced this page is, not how good Niche data refinery is. More in the glossary Self-reported Checked 4 days ago