Skip to content
RepoSniffer

RepoSniffer

There's a repo for that. Let RepoSniffer find it.
Get started GitHub

Semantic GitHub repo discovery for coding agents and humans. Describe a feature in plain language — “markdown editor with live preview” — and RepoSniffer returns a ranked, adoption-grade list of open-source projects that implement it, with evidence and quality signals (stars, activity, license, archived status).

Built to be the grounding layer for coding agents: stop hallucinating repos, get a verifiable “best of kind” answer with reasons.

Features

  • Semantic, not literal. GitHub search + grep.app match keywords; RepoSniffer matches intent via README embeddings.
  • Adoption-grade verdicts. Every result carries license classification, activity freshness, archived flags, and a recommendation — “good reusable project” beats “abandoned one”.
  • Zero-config embedding. Local ONNX model (bge-base-en-v1.5), no torch, no API key.
  • Pluggable. Swap the local model for any OpenAI-compatible embeddings API.
  • Fast repeats. Local SQLite cache of repos, READMEs, embeddings and query results.

Quickstart

Terminal window
# CLI
uvx reposniffer "markdown editor live preview" --top-k 5


# MCP server (stdio) — wire into opencode, Claude Code, Codex, Cursor, ...
uvx --from reposniffer reposniffer-mcp

Set GITHUB_TOKEN (or have the GitHub CLI authenticated) to raise search rate limits.

Quality

Retrieval quality is measured objectively against a golden eval set — currently hit@1 0.83, hit@3 0.83, hit@5 0.83. See Evaluation.