Skip to main content
fetchbean fronts a set of providers behind one key. You call any of their methods through run, passing the provider, its endpoint, and the input. fetchbean handles the keys, billing, and rate limits, and returns the provider’s native response. 
curl https://api.fetchbean.com/v1/run \
  -H "Authorization: Bearer $FETCHBEAN_KEY" \
  -H "Content-Type: application/json" \
  -d '{"provider":"exa","endpoint":"/search","input":{"query":"best vector databases","numResults":10}}'
The request body has three fields:
FieldDescription
providerThe provider id, for example exa, parallel, jina, tikhub, openweather, remoteok.
endpointThe provider’s own endpoint, for example /search or /v1/extract.
inputThe request body passed through to the provider.

​
A few examples

exa search
-d '{"provider":"exa","endpoint":"/search","input":{"query":"best vector databases"}}'
parallel search
-d '{"provider":"parallel","endpoint":"/v1/search","input":{"search_queries":["vector databases"]}}'
parallel extract
-d '{"provider":"parallel","endpoint":"/v1/extract","input":{"urls":["https://example.com"]}}'
jina read
-d '{"provider":"jina","endpoint":"/read","input":{"url":"https://example.com/post"}}'
Each method has a per-call credit cost and a defined set of parameters.

See the full catalog

Every provider and method, with parameters and per-call pricing.
A few common methods also have normalized shortcuts (such as POST /v1/search) that return a stable JSON shape. Those are a convenience layered on top of the same providers; run is the complete surface.