What it does
AI coding agents are very good until they are not. A requirement is unclear, a change is risky, a bug will not reproduce, or you just want a person to look over what the agent built. Connect humans fix ai and your agent gets one new ability: ask a human developer.
Your agent asks you first. Nothing is sent until you confirm by email.
The agent writes a short summary of the problem. We email you a code. You enter it, and the request goes to developers we have approved by hand. One of them calls you for a free video call, and most problems are solved there. If more work is needed, you get one fixed price and decide whether to accept it. You pay only after you approve the result.
Connect it
The connector lives at one address. You do not need an account or a key.
https://humansfix.ai/mcp
Claude Code
claude mcp add --transport http humansfix https://humansfix.ai/mcp
Codex
codex mcp add humansfix --url https://humansfix.ai/mcp
Cursor
Add this to ~/.cursor/mcp.json (or .cursor/mcp.json in one project), then restart Cursor.
{
"mcpServers": {
"humansfix": { "url": "https://humansfix.ai/mcp" }
}
}
Claude on the web
Open Settings, then Connectors, choose Add custom connector, paste the address above and give it a name.
Any other assistant that supports remote MCP servers can use the same address. Menus and commands change from time to time, so check your assistant's own help if a step looks different.
What happens when your agent uses it
- Your agent asks you. It should ask for your permission and for your email address before it does anything.
- It sends us a short summary of the problem and your email. It must not include passwords or keys, and we refuse text that looks like one.
- We email you a code. The request stays private, and no developer sees it, until you enter that code on our confirmation page.
- A developer proposes a time. You confirm it and join a free video call.
- The problem is solved on the call, or you get one fixed price. You accept it or say no, and you pay only after you approve the finished work.
What we keep, and what we never do
- We keep the summary and your email, like any request made on our site.
- We never see your code, your files or your conversation with the agent. We only get what the agent chooses to send.
- Nothing is published and nobody is contacted until you confirm.
- The connector never charges anything and never handles payment. If you pay for work later, it happens on our site, after you approve.
- We limit how many requests can be started, and how many emails one address can receive.
- If your email already has an account here, the agent tells you to sign in on the website instead.
For people building assistants
The connector is a stateless MCP server over Streamable HTTP. It needs no authentication and answers with JSON. It speaks the 2026-07-28 revision and also accepts the older initialize handshake from clients that use 2025-11-25, 2025-06-18, 2025-03-26 or 2024-11-05.
It has one tool, ask_human_developer, with these arguments:
summary(required): what the person needs, in plain words, 30 to 2000 characters. No secrets.email(required): the person's own email address.tool(optional): the AI tool the app was built with, orother.availability(optional): when the person is free for a call.
It starts a draft request and emails a confirmation code to the person. It does not publish anything, and errors come back as tool results with isError set, so the model can correct itself. A quick check from a terminal:
curl -s -X POST https://humansfix.ai/mcp -H 'Content-Type: application/json' -H 'Accept: application/json, text/event-stream' -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
Questions
Can my agent start a request without me knowing?
It can send us a summary, but nothing happens until you enter the code we email you. If you did not ask for it, ignore the email and the request is never sent to anyone.
Does it cost anything?
The first call is free. If more work is needed, a developer names one fixed price after seeing the problem. You can say no, and you pay only after you approve the finished work.
Does my agent see my payment details?
No. Payment never goes through the connector. It happens on our website, and only if you accept a price.
Which developers will see my request?
Only developers we have approved by hand, and only after you have confirmed it.
How do I remove the connector?
Remove the server in your assistant's settings. Requests you already confirmed stay on our site until they are finished.