gitcrawl is a local-first GitHub maintainer triage tool written in Go.
The target is a compact, local SQLite workflow for syncing, searching, clustering, and reviewing related GitHub issues and pull requests.
- local SQLite storage
- metadata-first GitHub sync for open issues and pull requests
- optional comment, review, review-comment, PR code, and PR review-thread hydration
- canonical thread document building
- bounded local Git source-document indexing
- FTS search
- OpenAI summaries and embeddings
- deterministic fingerprints
- vector search
- clustering and durable cluster governance
- portable sync export/import
- CLI JSON surfaces for automation and agents
- TUI browsing after core JSON contracts settle
- local HTTP API
- hosted service runtime
- browser web UI
- GitHub write-back actions
cmd/gitcrawl: executable entrypointinternal/cli: command parsing and outputinternal/config: config and env resolutioninternal/store: SQLite schema and persistenceinternal/github: GitHub API clientinternal/syncer: repository sync workflowsinternal/documents: canonical document generationinternal/codeindex: tracked source-file scanninginternal/openai: OpenAI summaries and embeddingsinternal/vector: vector search abstractioninternal/cluster: similarity and durable cluster governanceinternal/search: keyword, semantic, and hybrid searchinternal/portable: compact sync export/importinternal/tui: terminal UI
TUI guidance:
gitcrawl tui [owner/repo]is a supported command; omitowner/repoto use the most recently updated local repository- keyboard-first navigation is required
- mouse support is optional polish
- right-click must not be required for primary actions because terminal mouse support is inconsistent
- avoid decorative glyph noise or transient rendering debris in dense panes
No serve command.
Public commands:
initdoctorconfigureversionsyncrefreshsummarizekey-summariesembedclusterthreadsrunsclustersclusters-reportdurable-clusterscluster-detailcluster-explainneighborssearchcodeghclose-threadclose-clusterexclude-cluster-memberinclude-cluster-memberset-cluster-canonicalmerge-clusterssplit-clusterexport-syncimport-syncvalidate-syncportable-sizesync-statusoptimizetuicompletion
search also supports the common gh search read-only shape for cached discovery:
gitcrawl search issues <query> -R owner/repo --state open --json number,title,state,url,updatedAt,labels --limit 30
gitcrawl search prs <query> -R owner/repo --state open --json number,title,state,url,updatedAt,isDraft,author --limit 20
gitcrawl search issues <query> -R owner/repo --state open --sync-if-stale 5m --json number,title,url
This compatibility path reads from local SQLite by default. It avoids GitHub REST search quota and is not a replacement for final live gh verification before comments, closes, labels, or merges. --sync-if-stale <duration> may run one metadata sync first when the repository mirror is older than the requested max age; the search result itself still comes from SQLite.
gitcrawl gh moved to Octopool. Invoking it now prints a migration note:
gitcrawl gh moved to octopool.
Run: octopool login
Then use: octopool gh ... or symlink octopool as gh.
Octopool owns the org-authenticated shared GitHub CLI cache and pooled read relay. Gitcrawl keeps the local mirror/search/cluster/TUI product. Use gitcrawl search issues|prs ... for local discovery and octopool gh ... or a symlinked Octopool binary for pooled gh reads.
Gitcrawl portable stores carry repo snapshots only. They do not carry an Octopool or runtime gh cache.
Default config path:
~/.config/gitcrawl/config.toml
Default database path:
~/.config/gitcrawl/gitcrawl.db
Primary environment variables:
GITCRAWL_CONFIGGITHUB_TOKENOPENAI_API_KEYGITCRAWL_DB_PATHGITCRAWL_SUMMARY_MODELGITCRAWL_EMBED_MODEL
Legacy environment aliases may be supported only when they do not leak old naming into user-facing output.