Research & data
Open knowledge sources and search APIs for research agents.
Public-data and search connectors. Most require no credentials; a few use an API key.
Wikipedia (wikipedia) — no auth
| Operation | Purpose |
|---|---|
search | Search article titles. |
get-summary | Lead summary for an article. |
arXiv (arxiv) — no auth
| Operation | Purpose |
|---|---|
search | Search papers. |
get-paper | Metadata for a paper ID. |
save-pdf-to-files | Save the PDF to the user's files. |
Example: a literature-review agent calls search, then save-pdf-to-files for the top three results.
PubMed (pubmed) — no auth
| Operation | Purpose |
|---|---|
search | Search PubMed. |
get-abstract | Fetch an abstract. |
Open Library (openlibrary) — no auth
| Operation | Purpose |
|---|---|
search | Search books. |
get-work | Work metadata. |
Open-Meteo (open-meteo) — no auth
| Operation | Purpose |
|---|---|
forecast | Weather forecast for a location. |
geocode | Resolve a place name to coordinates. |
Exa (exa) — API key
Neural web search. Slug registered with authType: "api_key". The operation surface is loaded from the registry once an API key is configured.
Other knowledge sources
The registry also includes Coin Gecko (coin-gecko), OpenWeatherMap (openweathermap), Amadeus, Bitly, and many more. Browse /[orgSlug]/connectors for the full list and the per-connector operation set.