ripgrep for your team's knowledge base.
Hybrid retrieval, self-hosted, single command.
curl -fsSL https://ragrep.cc/install.sh | sh$ ragrep "how does the auth flow work"
Top 3 results for 'how does the auth flow work'
[1] [slack] #eng-platform — "new token rotation" rerank=0.89
We rotated service-to-service tokens to short-lived JWTs.
The auth middleware now validates via the shared signing key...
[2] [gdrive] Auth Architecture v2 — design doc rerank=0.84
OIDC login → IdP → callback with code → exchange for id_token
+ refresh. Refresh tokens live in an httpOnly cookie...
[3] [git] commit: migrate auth middleware to JWT rerank=0.81
Replaces session-cookie auth with signed JWTs. Rolls out
per-service behind a feature flag, fallback to cookies...Hybrid retrieval, not just embeddings. FAISS + BM25 + reranking, fused with RRF. Recalls cross-source content that pure dense or pure sparse misses.
Self-hosted, no lock-in. Your data stays on your machine. The index is three files you can cp or version-control.
Multi-source ingest. Slack, Confluence, Jira, Drive, Git, Bitbucket, local files — out of the box. Content-hash dedup means re-indexing only embeds what changed.
# after install
ragrep "how does the auth flow work"
ragrep "deploy" -m grep # exact substring
ragrep "incident" -s slack --after 2w # filter source + date
ragrep "auth" --json | jq # for agents/scripts