Build your knowledge base

Bootstrap and keep fresh

Run the whole pipeline in one command, compose the stages, and keep it fresh with cron or a git post-commit hook.

Rather than running each stage by hand, bootstrap chains them end to end, so a teammate goes from nothing to a fully-wired workspace in one command, and it's safe to re-run on a schedule to stay current.

bootstrap chains mirror repos → index → connect → embed → enrich → wiki → write editor steering, skipping anything not enabled:

contextlake bootstrap --llm builtin

--llm builtin powers the wiki stage with a zero-setup CPU model, so this single command builds the whole knowledge layer (graph, vectors, and wiki) for every repo. (--llm ollama | openai | auto for higher-quality prose; the pre-command form contextlake --llm builtin bootstrap also works.) Without any --llm, and with [llm] disabled in kb.toml, the wiki stage no-ops and the rest still runs. Because everything generated lives under one store_dir, setting it to a folder in your workspace keeps the entire knowledge base in a single, easy-to-access location.

Both config files are read from their default locations (~/.contextlake.ini and ~/.contextlake/kb.toml); pass --config / --kb-config to point elsewhere. The valid [kb] keys are store_dir, languages, skip_generated, max_file_bytes, and index_workers (plus the [embeddings], [llm], [sources], [rules] tables); an unrecognized key or table is warned and ignored, so a typo like store (for store_dir) is surfaced rather than silently dropping the run into the wrong place. Skip stages with --no-sync / --no-embed / --no-wiki / --no-connect / --no-enrich. For an isolated CLI, install with pipx install "contextlake[kb]", or run ad-hoc with uvx.

Command composition#

Every stage is standalone, idempotent, and composable. Use these flows to build exactly what you need:

contextlake bootstrap runs the full pipeline (mirror, index, connect, embed, enrich, wiki, steer) end to end, so init plus bootstrap takes a blank workspace to a mirrored, indexed, embedded, connector-enriched, wiki'd, editor-wired workspace in one command (skip enrich with --no-enrich).

Keep it fresh on a schedule#

bootstrap is incremental and branch-safe, so it's safe to run repeatedly, it re-mirrors, re-indexes only the repos whose HEAD moved, refreshes the knowledge layer, and rewrites the steering, without touching an in-progress working tree. Run it from cron:

*/30 * * * * contextlake bootstrap >> ~/.contextlake/refresh.log 2>&1

or as a systemd user timer, see examples/contextlake.service and examples/contextlake.timer.

Re-index on commit (git hook)#

For continuous freshness without a schedule, install a git post-commit hook that re-indexes a repo the moment you commit to it:

contextlake hook install                     # the repo in the current directory
contextlake hook install --workspace ~/src   # every git repo under a mirror
contextlake hook status  --workspace ~/src   # which repos are wired
contextlake hook uninstall                   # remove it (any pre-existing hook is kept)

The hook runs contextlake index <repo> detached (so the commit returns immediately) and re-uses the repo's stored id, so it updates the same graph node rather than a duplicate. Mirror-wide syncing (fetch new clones, prune) still belongs to bootstrap on a schedule; the hook keeps local edits current between syncs.

If two contextlake processes ever target one store at once (say a bootstrap and a hook-triggered index), the second takes an advisory single-writer lock (<store_dir>/.contextlake.lock) and refuses rather than interleaving SQLite writes, naming the process that holds it. A lock left by a crashed run is reclaimed automatically; override (rarely correct) with CONTEXTLAKE_ALLOW_CONCURRENT=1.

See also#

• Index the code graph
• Generate the wiki
• Reading the console output