Who it is forA knowledge worker who writes notes all day and wants answers they can trust before acting on them.

The problemNew notes are not searchable until reindexed, and answers can state claims without evidence.

The outcomeChanged documents stay indexed, quick memories are searchable by tag, and answers can be verified claim by claim.

What you need first

Python 3.11 or newer
git and a terminal
An embedding endpoint that supports /embeddings, plus a chat endpoint (may be a local one)
A notes or project docs directory to index

Steps

Run loci watch in a separate terminal to keep changed documents indexed.
Run loci remember "Use staging for demos" --title "Demo environment" to save a quick memory.
Run loci search "demo environment" --tag memory to confirm the memory is found.
Run loci ask "How should I prepare a demo?" --verify to get an answer with claim-by-claim verification.

Prompt for your agent

Using the IvenKooLab/loci repo, run loci watch in <instance-path>, then save a memory with loci remember and ask <question> using loci ask with --verify. Expected output: a cited answer with verification results. Verify that the memory saved with loci remember appears in loci search results filtered by --tag memory.

Caveat. The README says that after changing the embedding model you must rebuild with Loci ingest --force, or the existing index will not match.

Use this in your AI agent

Ask your agent

Using the Deduped MCP, get the setup steps for IvenKooLab/loci and "Keep notes indexed and check answers before you rely on them"

Add Deduped to Claude Code once

claude mcp add --transport http deduped https://dedupedai.com/mcp

Cursor, ChatGPT, others

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer

What you need first

Steps

Prompt for your agent

Use this in your AI agent

Run the default town and inspect it in the dashboard

Compile your notes into a searchable Markdown wiki

Save a Bilibili or YouTube video as Markdown notes