myAgents
Who it's forPlaybooksWhy it worksShowcaseGuidesPlatformPricingmyOS ↗Sign inGet started
Start free
myAgents for…
Solo entrepreneursFreelancersSmall business ownersExecutives & foundersProductivity NerdsParentsTravelersYouTubersTikTok & Reels creatorsPodcastersWritersCoaches
Take your AI team anywhere
Now on the App Store for iPhone & iPad — and on Google Play for Android.
Download on theApp StoreGET IT ONGoogle Play
myAgents·Part of the myOS family
Why it worksShowcasePlaybooksGuidesConnectorsPlatformMusePricingSupportTermsPrivacyContact© 2026 myOS Inc.
Guides/Product

Connect Claude, ChatGPT, Gemini, or Muse to Your Agent Team

Your favorite AI chat is great at thinking and terrible at doing. Connect it to your agent team and the fifty-message brainstorm becomes six assigned tasks.

6 min read·Updated October 3, 2026

The short version: the AI chat you already pay for can now sit on top of your agent team. You think out loud in Claude, ChatGPT, Gemini, or Meta Muse. When the thinking turns into tasks, your assistant creates the work items, assigns the right agents, and your team gets to work in myAgents. The brainstorm is covered by your chat subscription. Your credits go to the real work in myAgents: each agent run, every one itemized, plus a small drafting cost when you design a fully custom team.

Promote yourself to boss. Hire your first AI agent in about two minutes. $10 in free credits, no card. It drafts, you approve.

Start free →

Why split thinking from doing

An hour of planning can be fifty messages. The execution it produces might be six tasks. If both happen in the same metered place, you pay execution prices for conversation. Connect your chat to your team and each side does what it's best at: the chat holds the long, messy conversation, the agents hold the jobs, the schedules, and the activity log.

It also works in reverse. Once connected, your assistant can see your team. Ask it "what are my agents up to?" over morning coffee and it reads the queue instead of guessing.

What a connected assistant can do

  • See the team. Agents, work queues, schedules, deliveries, spend.
  • Hand off work. Create work items, assign agents, comment with @mentions that wake the right agent, change statuses.
  • Send files. Attach the brief, the screenshot, the CSV right to the task. Your agent reads it on its first run.
  • Not fake anything. Your assistant acts as you, with your permissions, and the rules don't bend for it. It cannot mark work delivered. Only an agent that actually ran and produced something real can do that.

Connect Claude

On claude.ai (web or mobile): Settings, then Connectors, then add a custom connector with your server address:

`` https://www.myagents.ca/api/mcp/mcp ``

Claude sends you to a consent page where you sign in and choose what to allow: read only, or read and act. Approve and you're connected.

Using Claude Code? It's one command, and the Connect your AI tab in your team settings hands you the exact snippet with a fresh token baked in. Claude Desktop works too, via a small bridge; that snippet is in the same tab.

Connect ChatGPT

ChatGPT calls these custom connectors, and they live behind developer mode (paid plans, on the web). Turn it on under Settings, then Security and login. Then open Plugins, tap +, choose a public endpoint and paste the same server address. ChatGPT runs the same sign-in and consent flow, and the same read or act choice is yours to make.

Connect Gemini

The Gemini app can't add custom connectors yet, but the Gemini CLI can, and it needs no token at all. Add the server to your settings file:

``json { "mcpServers": { "myagents": { "httpUrl": "https://www.myagents.ca/api/mcp/mcp" } } } ``

On first use the CLI notices the server wants sign-in, opens your browser to the same consent page the others use, and you pick read or read and act right there. Prefer a token anyway? Mint one in the Connect your AI tab and add it as an Authorization header:

````

Try it on your own work

Hire your first team in about two minutes. $10 in free credits, no credit card required.

Start free →See pricing

Frequently asked questions

Does connecting Claude, ChatGPT, Gemini, or Muse cost extra?
The connection itself is free. When your assistant assigns work and an agent runs, that run costs credits exactly like it would if you'd assigned it in the app. Approving a ready-made template team assembles free; approving a fully custom design, or adding a special hire, spends a small amount of credits to draft those teammates. Reading your team's status costs nothing.
Which assistant should I connect?
The one you already pay for and talk to every day. The server speaks an open standard (MCP), so Claude, ChatGPT, Gemini, and Meta Muse all connect to the same address and get the same abilities. Claude, ChatGPT, and Gemini sign you in on a consent page; Muse takes a token in one pasted message. There's no wrong pick.
Can my assistant do things I couldn't do myself?
No. It acts through your account with your permissions, and the platform's rules apply to it the same way they apply in the app. It can hand your team work, but it can't mark that work delivered. Only an agent that actually ran can produce a delivery.
What if I only want my assistant to watch, not act?
Choose read only when you approve the connection, or mint a read-only token. Your assistant can then see agents, queues, and deliveries. The one thing it can save is a proposed team design, stored as a private draft that only you can turn into a team. Beyond that it can't create or change anything, and you can revoke any token at any time.

See the team built for you

myAgents ships ready-made teams tuned to how you work:

For Solo entrepreneurs →For Small business owners →For Freelancers →

Related guides

All guides →
Product
How to Share What Your AI Knows With Your Team
Connect Claude
How to Connect Claude to Gmail
Connect Claude
How to Connect Claude to Google Calendar
Connect Claude
How to Connect Claude to GitHub
json { "mcpServers": { "myagents": { "httpUrl": "https://www.myagents.ca/api/mcp/mcp", "headers": { "Authorization": "Bearer <your token>" } } } }

That's it. Your Gemini sessions can now see the team and hand off work like the others.

Connect Muse

Meta Muse builds custom connectors from a plain description, so connecting it is one message. Mint a token in the Connect your AI tab in your team settings, then paste this into a new Muse chat with your token in place:

`` Build a custom integration to myAgents, my AI team. Its MCP server is https://www.myagents.ca/api/mcp/mcp (Streamable HTTP over POST; accept both application/json and text/event-stream responses). Every request needs the header "Authorization: Bearer <paste your token here>". List its tools, then save this as a reusable skill called "myAgents". Use it whenever I ask you to hand work to my team, check on my agents, read a delivery, or hire a team. ``

Muse builds the connector, lists the tools, and keeps it as a skill. You never have to type the token in by hand: the Connect your AI tab hands you this exact message with your new token already in it, and so does the connect screen right after sign-up, where a new account gets its Muse token in one click. Muse can then interview you and propose your first team for you to approve.

Read only lets Muse watch and propose. Read and act lets it hand out work. The full walkthrough, including the one line that teaches Muse when to delegate, is in the Muse connect guide.

The same rules apply to all of them

A connection is a credential, and you hold the leash. Mint a read-only token when you want a window, not a lever. Every agent run your assistant kicks off is itemized like any other run, and surprise bills stay impossible. Revoke a token and the connection dies on the spot.

One honest note: a connected assistant sees what you'd see in the app, your team's status, work, and deliveries. Connect the assistants you trust, and give read-only to the rest.

The Connect your AI tab lives in your team settings. New here with no team yet? The connect screen greets you right after sign-up.

You start with $10 in free credits, then from $25 a month. The conversation with your assistant costs whatever your chat plan costs. The work your team does is the part we meter, because that's the part that's real.