| name | freesearch-search |
| description | Run the lit-review orchestrator free index search channel against keyless OpenAlex, Crossref, and Semantic Scholar keyword-search endpoints. Use when the user invokes the free search channel, asks for keyless scholarly index search, or needs Stage 4e fallback discovery without SearchAPI, Gemini, or Undermind credentials. |
Free Index Search channel (Stage 4e) — keyless OpenAlex / Crossref / Semantic Scholar
A zero-key discovery channel: it runs the Stage-0 query list against the free,
keyless keyword-search endpoints of OpenAlex, Crossref, and Semantic Scholar and
normalizes the real records into the pipeline schema. Together with the agent
web-search channel (Stage 4d) it is the "only Claude Code" search fallback — no
search account or API key required.
Unlike web search, this is a plain subprocess (it calls keyless HTTP APIs), so it
runs the same way in any context.
When to use
- The user has no
SEARCHAPI_API_KEY / GEMINI_API_KEY / Undermind login — this is
the keyless search backbone, OR
- You want extra grounded, clean-metadata coverage from the scholarly indexes
alongside the other channels.
Run it in parallel with whatever else is available; its output merges at dedup.
Recipe
Use the same Stage-0 query list Google Scholar uses:
python freesearch-search/scripts/freesearch_search.py \
--queries-file OUT/scholar_queries.json -o OUT/stage4e_freesearch.json
Or a single query with --query "...". Options: --per-query N (default 20) and
--sources openalex,crossref,semanticscholar (any subset). It writes
stage4e_freesearch.json (+ .ris) with source set per index, deduped across
indexes by DOI then title. The stage[0-9]*.json dedup glob picks it up
automatically.
Keys
None required. OpenAlex and Crossref use the polite pool via a mailto (set
LITREVIEW_CONTACT_EMAIL to use your own contact). If present, OPENALEX_API_KEY
and SEMANTIC_SCHOLAR_API_KEY are used only to raise rate limits. It makes no LLM
API calls.
Anti-hallucination
Every hit is a real index record, so there is nothing to fabricate — relevance, not
existence, is the only variable, and screening handles relevance. Stage 5b
verification still confirms each paper downstream. The keyless Semantic Scholar tier
is rate-limited, so it is queried with a short delay and skipped gracefully on a 429.
