corvid-remember 0.2.0

Permanent memory layer for AI coding agents

corvid

A permanent memory layer for AI coding agents. Installs as a native skill. One command: /remember.

Inspired by Karpathy's llm-wiki pattern for building knowledge bases with LLMs, but instead of ingesting external docs, corvid captures what you learn during the conversation itself.

Why

Every AI session starts from zero. You figure something out, close the terminal, and next month your agent has no memory of it.

Built-in memory is per-project, 200 lines, no search. corvid gives you cross-project memory with semantic search. What you figure out once stays figured out.

Install

pip install corvid-remember
corvid install

corvid install detects your agents (Claude Code, Codex, Gemini CLI) and adds /remember as a native skill in each one.

$ corvid install

  ✓ Claude Code      → ~/.claude/skills/remember/SKILL.md
  ✓ Codex CLI        → ~/.codex/skills/remember/SKILL.md

  Wiki: ~/corvid
  Semantic search: enabled (BAAI/bge-small-en-v1.5)

Lightweight mode (keyword search only, no embedding model):

pip install corvid-remember[lite]
corvid install

How it works

Say /remember and your agent distills the current insight into a searchable article:

You: /remember the Google OAuth deploy gotchas

corvid writes → ~/corvid/wiki/auth/google-oauth-deploy-gotchas.md

  # Google OAuth Deploy Issues
  Three things break after deploy that work locally:
  1. Callback URL must match EXACT casing in Google Console
  2. Consent screen redirect URI needs the production domain with https
  3. Cookie SameSite=None + Secure required behind reverse proxy
  Symptoms: silent 400 on redirect, no error in server logs.

Next session, any project, your agent pulls up what you already solved.

Search

Two search modes, both local, both fast.

Keyword (always on): SQLite FTS5 with Porter stemming and BM25 ranking. "Finetuning" matches "fine-tuned." Zero dependencies.

Semantic (full install): fastembed generates embeddings with a 33MB model (bge-small-en-v1.5), sqlite-vec stores and searches them inside the same SQLite database. "Login callback issue" finds your article about "OAuth redirect" even though it never uses those words. Runs on CPU in ~50ms. No GPU. No server. No API calls.

Both modes run together. Keyword results first, then semantic fills in what keyword missed.

corvid search "oauth redirect"
corvid search "liability caps" --json
corvid stats

Built-in memory vs corvid

	Built-in	corvid
Scope	Per-project	Cross-project
Limit	~200 lines	Unlimited
Search	None	Keyword + semantic
Control	Agent decides	You decide with /remember

What people save

• The auth fix that works locally but breaks after deploy
• The contract clause taxonomy the parser needs to handle
• The database index that stopped the timeout
• The API behavior that is not in the docs
• The architecture decision and why you made it
• The deploy config that makes CI pass after it randomly started failing

What gets saved

Your agent writes markdown articles by category. You never touch the structure.

~/corvid/
  corvid.db          # search index (disposable, rebuildable)
  INDEX.md           # table of contents your agent maintains
  wiki/
    auth/
      google-oauth-deploy-gotchas.md
    contracts/
      liability-cap-types.md
    backend/
      supabase-rls-service-role.md

Real articles with tables, exact values, commands, reasoning. Not chat logs.

Under the hood

One Python file. SQLite FTS5 for keyword search. sqlite-vec for vector search. fastembed for embeddings (ONNX, CPU-only, 33MB). Everything local. Database is disposable, rebuild anytime with corvid index-all.

Commands

Command	What it does
corvid install	Detect agents, install /remember skill, init database
corvid search <query>	Search (human-readable)
corvid search <query> --json	Search (JSON for agents)
corvid index <file>	Index one markdown file
corvid index-all	Re-index everything
corvid stats	Show article counts

License

MIT