Skip to content

v0.3.1

Choose a tag to compare

View all tags
@fatwang2 fatwang2 released this 22 Oct 11:05
· 25 commits to main since this release
v0.3.1
f77a157
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Install from npm:

npm install search2ai@latest

Existing Exa, Tavily, Search1API, SearXNG and Firecrawl clients can now use dedicated API bases for supported search and single-page reading calls, while upstream providers remain independently configurable.

  • SearXNG JSON search: root and /search, GET and form POST.
  • Exa: /search and /contents.
  • Tavily: /search and /extract.
  • Search1API: /search, /news and single-page /crawl.
  • Firecrawl: /v1/scrape and /v2/scrape for one explicit URL.

Set the client's API base to https://YOUR_GATEWAY/compat/PROTOCOL and use your gateway key. Supported calls keep their existing SDK methods and response parsing. Matching upstreams preserve native fields; cross-provider conversion supports the documented subset and rejects options it cannot fulfill. This is not full vendor API compatibility. Whole-site crawling, job APIs and vendor-specific MCP compatibility are outside scope. MCP continues to expose unified gateway tools.

The existing /v1/* interface remains compatible. This release also fixes repeated crawl fallback attempts, shares cancellation and empty-result policy, bounds compatibility requests and responses, and updates CORS and OpenAPI routes.

English is now the default README, with Chinese documentation available. See the compatibility matrix and SDK examples (中文) for exact limits.

Validation: typecheck, build, 86 tests, 18 pinned SDK/integration HTTP calls across native and translated modes, and a Workers dry-run bundle passed. The packaged JS artifact also passed the 18 SDK calls. Upstreams were mocked; live search quality was not tested.

Implemented in #50.

Assets 4
Loading