Similarity search
agentmine similar is Agentmine’s main entry point for finding prior work.
Usage
agentmine similar "React Router auth redirect loop" --limit 5agentmine similar "schema migration" --source codexagentmine similar "test flake timeout" --project /path/to/repoagentmine similar "today's agentic docs work" --root-only --since 1dAuto mode
By default, similar runs in auto mode:
- It uses FTS when no local embedding index is available.
- It can use hybrid search when local embeddings exist and guardrails are satisfied.
- It returns reconstruction commands such as
agentmine session <id> --md. - It excludes runtime-injected instructions, skill payloads, hook feedback, compaction handoffs, and automatic approval transcripts from matching content.
Pass --include-injected only for corpus diagnostics. It restores those messages to both lexical
and semantic candidate searches and preserves raw injected session titles; otherwise those titles
are returned as null.
Time and lineage filters
Use --root-only to exclude normalized child workers and automatic reviewers. Combine it with
--since and --until for a bounded investigation:
agentmine similar "agent-context-kit agentic docs" \ --all-projects \ --root-only \ --since "2026-07-23T00:00:00+03:00" \ --until "2026-07-24T00:00:00+03:00"--since is inclusive and --until is exclusive for ISO timestamps. A bare YYYY-MM-DD
--until includes that complete UTC day, matching the shared Agentmine date-filter contract.
Impossible calendar dates are rejected instead of rolling into a later month. When the user’s day
is not UTC, pass explicit ISO offsets as shown above.
Optional local embeddings
Local embeddings are optional. To enable semantic and hybrid search, pull an embedding model with Ollama and build the local index:
ollama pull nomic-embed-textagentmine embed --provider ollama --model nomic-embed-text --dry-runagentmine embed --provider ollama --model nomic-embed-text --limit 500agentmine similar "agent first CLI JSON stdout stderr" --mode hybridembed --dry-run plans the run without writing chunks, vectors, or run receipts. Once you have
run it without --dry-run, similar can select hybrid mode automatically, or you can request it
explicitly with --mode hybrid (or --mode embedding for pure semantic retrieval).
Scope an embedding run to recent work with --since — it accepts an ISO date (YYYY-MM-DD) or a
relative offset such as 7d, 2w, or 12h, and filters both planning and the pending-chunk
queries by session start time, so only chunks from sessions in the window get embedded.