Skip to main content
NeuronFeed
Integrations

Plug the directory into your stack

All directory data — 2,700+ startups, funding rounds, scores — is consumable without an account. Pick your route:

Building with AI agents?

The directory speaks MCP natively at https://neuronfeed.com/api/mcp — add it to Claude or any MCP client and query startups conversationally. Agents can also submit startups programmatically.

Frequently asked questions

Do I need an API key to use the NeuronFeed API?

No. Every read endpoint is public — no key, no OAuth, no signup. CORS is open on GET requests, so browser-side calls work too. Public paths all live under /api/v1/, /api/mcp, and /api/tools/; if you ever get a 401 you have hit an admin route by mistake.

What are the rate limits?

The REST API allows 60 requests per minute per IP. Responses carry X-RateLimit-Limit and X-RateLimit-Remaining headers, and a 429 includes Retry-After so you can back off cleanly. AI-powered tool endpoints such as Visual TL;DR have their own daily quotas, documented on each tool page.

Can AI agents access the directory data?

Yes. The directory speaks MCP natively at https://neuronfeed.com/api/mcp, so you can add it to Claude or any MCP client and query startups conversationally without writing integration code. Agents can also submit startups programmatically through the agent submission endpoint.

Is the API free to use commercially?

The data is free to read and there is no paid tier for access. Attribution back to NeuronFeed is expected if you republish records, and the 60 req/min limit applies to everyone — if your use case needs more headroom, get in touch before building around it.

How often does the data change, and how should I cache it?

Directory data updates on a daily cadence rather than in real time, so cache list responses for an hour or more. For "what is new" workflows, poll once or twice a day with ?sort=newest and stop paging as soon as you hit a slug you have already processed.

Can I connect it without writing code?

Yes. Because the endpoints are plain unauthenticated JSON over HTTPS, a single HTTP module is enough in most no-code tools. We publish step-by-step recipes for Make and for Zapier, including field mapping and how to trigger on new startups or new funding rounds.