AIDataParser MCP server
MCP serverEverything elseExtract schema-guaranteed JSON from PDFs, images and messy text. 50 free credits, no card.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use AIDataParser MCP server
Install AIDataParser MCP server
The server’s own address, for the clients that take one directly. Or connect ahel onceand every client you use reads it from one address, with the account kept on ahel rather than in each client’s config.
Claude Code
claude mcp add --transport http aidataparser-mcp-server 'https://aidataparser.com/v1/mcp'Run it once in your project, then open /mcp to approve any sign-in the server asks for.
Claude Desktop
https://aidataparser.com/v1/mcpAdd a custom connector in Settings, paste this address, and approve the sign-in.
Cursor
cursor://anysphere.cursor-deeplink/mcp/install?name=aidataparser-mcp-server&config=eyJ1cmwiOiJodHRwczovL2FpZGF0YXBhcnNlci5jb20vdjEvbWNwIn0=Open the link and Cursor adds the server at that address.
ChatGPT
https://aidataparser.com/v1/mcpIn Settings, enable Developer mode, create an MCP app, and paste this address. Your plan and workspace must allow custom apps.
Codex
codex mcp add aidataparser-mcp-server --url 'https://aidataparser.com/v1/mcp'Run it once, then sign in with codex mcp login aidataparser-mcp-server if the server asks for an account.
From the project's README
As published by shibley/aidataparser-mcp-server in README.md.
The official Model Context Protocol registry record for AIDataParser — a credit-based data-extraction API that turns PDFs, images and messy text into schema-guaranteed JSON.
There is no server code in this repo. The MCP server is hosted: it lives at
https://aidataparser.com/v1/mcp and speaks JSON-RPC 2.0 over Streamable HTTP.
This repo holds the server.json that is the source of our record in
registry.modelcontextprotocol.io,
the canonical upstream every downstream MCP catalog syncs from.
Add it to your client
{
"mcpServers": {
"aidataparser": {
"type": "http",
"url": "https://aidataparser.com/v1/mcp",
"headers": { "Authorization": "Bearer adp_live_..." }
}
}
}
Get a key (50 free credits, no card):
curl -X POST https://aidataparser.com/v1/keys \
-H 'Content-Type: application/json' \
-d '{"email":"you@example.com"}'
list_schemas and validate are free and need no key, so a client can add
the server and get something useful back before signing up.
Tools
| Tool | Cost | What it does |
|---|---|---|
parse_document | 1 credit | PDF or image (url or base64) → JSON matching your schema |
parse_text | 1 credit | Raw/messy text → JSON matching your schema |
infer_schema | 1 credit | One sample → a reusable JSON Schema |
validate | free, no key | Check a JSON value against a schema |
check_credits | free | Credits remaining on the key |
list_schemas | free, no key | Built-in templates (invoice, receipt, resume, …) |
Credit packs: https://aidataparser.com/pricing
Publishing
.github/workflows/publish-mcp.yml publishes on a push to main using GitHub
OIDC — no secrets. Bump version in server.json to cut a release; a push that
does not change the version is a no-op.
Before it publishes, the workflow runs two gates:
npm test— static checks onserver.json(validate-server-json.mjs). Each rule maps to a real failure from a sibling repo, e.g. the registry's 100-char description cap that 422'daisotools-mcp-serverv0.1.0.npm run check:remote— a liveinitialize+tools/listhandshake against the URL about to be advertised. A registry record pointing at a dead endpoint is worse than no record.
Both are runnable locally.
Advanced
- Delivery
- aidataparser MCP server → your ahel gateway (mcp.ahel.ai) → every connected AI client.
- Catalog kind
- mcp-server
- Gateway key
io-github-shibley-aidataparser- Source
- github.com/shibley/aidataparser-mcp-server
- Hosted endpoint
https://aidataparser.com/v1/mcp