Serveur MCP AI Form API

AI Form API: MCP servers

AI Form API is a forms API and MCP server. Your AI agent creates a form, you share the link, people answer, and the agent reads the answers as JSON or CSV.

This repo holds the connection docs, the server.json files for the official MCP Registry and code examples. The service runs at https://aiformapi.com. This repo has no server code.

• Docs: https://aiformapi.com/docs
• MCP setup page: https://aiformapi.com/docs/mcp
• OpenAPI 3.1 spec: https://aiformapi.com/openapi.json
• Full reference for agents: https://aiformapi.com/llms-full.txt
• Contact: hello@aiformapi.com

The 2 MCP servers

Both servers use the Streamable HTTP transport.

Server	URL	Auth	Who uses it
Creator	https://aiformapi.com/mcp	Authorization: Bearer afk_...	You, the form owner. Create forms and read responses.
Fill	https://aiformapi.com/mcp/fill	None. The fill code is an argument.	A person who hands a form to their own AI agent.

Get an API key for the creator server from the dashboard at https://aiformapi.com. See Authentication. The key starts with afk_. Keep it out of files you commit.

The fill server needs no login. A visitor presses "Want your AI agent to fill this?" on a public form page and copies a prompt with a one-time fill code. The agent reads the form, asks the person for anything missing, shows all answers, and submits only after the person says yes.

Which apps work today

App	Creator server	Fill server
Claude Code	Yes, with a header	Yes
Cursor	Yes, with a header	Yes
VS Code	Yes, with a header	Yes
Claude Desktop	Yes, through mcp-remote with a header	Yes, as a custom connector
n8n	Yes, with a header	Yes
Claude.ai	Not yet. It needs OAuth, which is planned.	Yes, as a custom connector
ChatGPT	Not yet. It needs OAuth, which is planned.	Yes, as a custom connector

Connect the creator server

Set your key in your shell first. Use your own key.

export AIFORMAPI_KEY="afk_paste_your_key_here"

Claude Code

claude mcp add --transport http aiformapi https://aiformapi.com/mcp \
  --header "Authorization: Bearer $AIFORMAPI_KEY"

This stores the key in your local Claude Code config. To keep the key out of a shared file, use a project .mcp.json that reads the environment variable (example):

{
  "mcpServers": {
    "aiformapi": {
      "type": "http",
      "url": "https://aiformapi.com/mcp",
      "headers": {
        "Authorization": "Bearer ${AIFORMAPI_KEY}"
      }
    }
  }
}

Run /mcp in Claude Code to see the connection. Claude Code docs: https://code.claude.com/docs/en/mcp

Cursor

Put this in .cursor/mcp.json for one project, or in ~/.cursor/mcp.json for all projects (example). Set AIFORMAPI_KEY in your shell profile first.

{
  "mcpServers": {
    "aiformapi": {
      "url": "https://aiformapi.com/mcp",
      "headers": {
        "Authorization": "Bearer ${env:AIFORMAPI_KEY}"
      }
    }
  }
}

Cursor docs: https://cursor.com/docs/context/mcp

VS Code

Put this in .vscode/mcp.json, or run MCP: Open User Configuration for a profile-wide file (example). VS Code uses the top-level key servers, not mcpServers. It asks for the key when the server starts for the first time.

{
  "inputs": [
    {
      "type": "promptString",
      "id": "aiformapi-key",
      "description": "AI Form API key",
      "password": true
    }
  ],
  "servers": {
    "aiformapi": {
      "type": "http",
      "url": "https://aiformapi.com/mcp",
      "headers": {
        "Authorization": "Bearer ${input:aiformapi-key}"
      }
    }
  }
}

VS Code docs: https://code.visualstudio.com/docs/agents/reference/mcp-configuration

Claude Desktop (through mcp-remote)

Claude Desktop reads local servers from claude_desktop_config.json. The mcp-remote package turns a remote server with a header into a local one. You need Node.js.

1. Open the file. On macOS it is ~/Library/Application Support/Claude/claude_desktop_config.json. On Windows it is %APPDATA%\Claude\claude_desktop_config.json. You can also open it from Settings, Developer, Edit Config.
2. Add the server (example). The space of Bearer lives in the environment variable, not in args, so the entry also works on Windows.

{
  "mcpServers": {
    "aiformapi": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://aiformapi.com/mcp",
        "--header",
        "Authorization:${AIFORMAPI_AUTH}"
      ],
      "env": {
        "AIFORMAPI_AUTH": "Bearer afk_your_key_here"
      }
    }
  }
}

1. Quit Claude Desktop fully and open it again.

This file holds your key in plain text. Do not share it or commit it.

n8n

Use the MCP Client Tool node and connect it to an AI Agent node. Set the endpoint to https://aiformapi.com/mcp, the transport to HTTP Streamable, and the authentication to Bearer with your key. Field names change between n8n versions, so check yours. n8n docs: https://docs.n8n.io/integrations/builtin/cluster-nodes/sub-nodes/n8n-nodes-langchain.toolmcp/

Connect the fill server

The fill server has no login, so it needs no header.

Claude.ai and Claude Desktop

Add it as a custom connector.

1. Go to Customize, then Connectors.
2. Press + Add, then Add custom connector.
3. Enter a name, for example AI Form API fill.
4. Enter the URL https://aiformapi.com/mcp/fill.
5. Choose No sign in for authentication.
6. Press Add.
7. In a chat, press +, then Connectors, and switch the connector on.

Anthropic says custom connectors are available on Free, Pro, Max, Team and Enterprise plans, and that Free plans have a limit of 1 custom connector. See https://support.claude.com/en/articles/11175166-get-started-with-custom-connectors-using-remote-mcp

ChatGPT

ChatGPT adds remote servers as connectors in developer mode. Create a connector, enter the URL https://aiformapi.com/mcp/fill, and choose no authentication. Menu names and plan availability change often. Check OpenAI's current help page for the exact steps.

Claude Code, Cursor, VS Code and n8n

Add it the same way as the creator server, without the header.

claude mcp add --transport http aiformapi-fill https://aiformapi.com/mcp/fill

{
  "mcpServers": {
    "aiformapi-fill": {
      "url": "https://aiformapi.com/mcp/fill"
    }
  }
}

Tools

Creator server (/mcp)

The tools return rows and form definitions. They do not summarise. Your own model reads the rows and writes the summary.

Tool	Arguments	Same as
create_form	title, description (optional), fields, settings (optional)	POST /v1/forms
list_forms	limit (optional), cursor (optional)	GET /v1/forms
get_form	form_id	GET /v1/forms/{id}
update_form	form_id, and any of title, description, fields, settings	PATCH /v1/forms/{id}
close_form	form_id	POST /v1/forms/{id}/close
reopen_form	form_id	POST /v1/forms/{id}/reopen
get_responses	form_id, and any of format, since, limit, cursor, order	GET /v1/forms/{id}/responses
add_webhook	form_id, url	POST /v1/forms/{id}/webhooks
get_usage	none	GET /v1/usage

The arguments match the body or query of the REST call. The form id is the argument form_id, not a path part. format is json (default) or csv. order is asc or desc. In update_form, fields replaces the whole list and settings merges key by key. There is no tool to delete a form. Use the REST API or the dashboard.

A failed call returns a tool error with the same JSON error body as the REST API. See API error codes.

Fill server (/mcp/fill)

Tool	Arguments	Does
get_fill_form	fill_token	Returns the form and instructions. Does not use up the code.
submit_fill	fill_token, answers	Stores the answers. answers maps field ids to values.

fill_token is the code from the prompt. It starts with fil_. A successful submit_fill returns response_id, receipt and submitted_at.

Limits on the free plan

Scope	Free plan
API key (REST and creator MCP together)	30 requests per minute, 1,000 per day
Open forms per account	3
Responses per account per calendar month (UTC), all sources	100
Fill MCP calls, per IP	20 per minute
One fill code	1 successful use, 60 minutes lifetime, 5 POST attempts
POST /v1/fill/{token}, per IP	10 per minute
Webhooks per form	3
Request body size	64 KB

Creator MCP calls count against the same limit as REST calls for the key. The API, MCP, agent fill and CSV export are included in the free plan. A 429 response has a Retry-After header. The full table, with the RateLimit-* headers, is at https://aiformapi.com/docs/rate-limits

Examples

The examples/ folder has copy-paste code from the quickstart. Each example creates a small RSVP form and reads the responses back.

• examples/curl/create-and-read.sh
• examples/javascript/create-and-read.mjs (Node 18 or newer, fetch)
• examples/python/create_and_read.py (requests)
• examples/mcp/: the client config files from this README

MCP Registry

The two servers are listed in the official MCP Registry with the server.json files in this repo:

• server.json: creator server, io.github.sum-eet/aiformapi
• server-fill.json: fill server, io.github.sum-eet/aiformapi-fill

License

The docs and examples in this repo use the MIT license. AI Form API is a hosted service. This license does not cover the service.