Memory Ingest
Claude Code maintains its own auto-memory — markdown it distills for itself — under:<encoded-cwd> is the repo’s absolute path with every character other than a
letter or digit replaced by - (e.g. /home/you/code/my_app →
-home-you-code-my-app). Claude Code cuts a name longer than 200 characters to
200 and ends it with a short hash.
claude-mem memory ingest imports those topic files directly into your
memory database as observations.
Why it doesn’t cost anything
Auto-memory is already distilled — each topic file is the same kind of artifact the observation generator produces. So memory-ingest does not run the Haiku generation pipeline. It stores each file’s prose directly through claude-mem’s existing observation seam (content-hash dedup + Chroma sync). Re-running generation on already-distilled prose would be lossy and pay for negative value. TheMEMORY.md index is skipped — it’s just links and carries no knowledge of
its own.
Usage
Always dry-run first — it’s a zero-spend, DB-free scan + count:Flags
How it runs
--dry-runis pure parse + count — it runs entirely in the CLI process, touches no worker and no database, and spends nothing.- The real ingest stores into the SQLite observation database, which lives
in the worker. The CLI starts the worker if needed and drives the import over
HTTP (
POST /api/memory/ingest), as summaries reach the worker. Bulk imports are not time-limited. - A file that cannot be read is reported as
failed; the rest of the import carries on. A symlinked note is never followed and a note over 64 KB is not read; both are reported asskipped, each named with its reason. A note swapped for a symlink mid-scan is refused when opened and reported asfailed. A source outside the projects directory, or one that does not exist, is a400from the worker. - Only the memory notes are imported. A sibling transcript is read for its
cwd, to key the project like live capture does; its content is never ingested.
Frontmatter
Topic files carry a small YAML frontmatter block, which is preserved as observation metadata:metadata.type (e.g. project, feedback, reference, user) and
metadata.originSessionId are carried through; files without frontmatter are
stored using their body as-is.
Idempotency
Ingest is safe to re-run. Observations are content-hash deduplicated on insert, so already-imported files are reported asalready-imported and skipped — only
new or changed files are stored.
Changed prose is stored even when the file keeps the same title and
CLAUDE_MEM_DEDUP_ENABLED=true. Native notes use exact content deduplication;
the title-only merge for generated observations does not discard note revisions
or distinct topic files sharing a title. Earlier revisions remain available.
The summary line reports the outcome:
Related
- Memory Export/Import — share memory sets between installations.