agi-memory
Your AI coding assistant forgets everything between sessions. This remembers.
Decisions you already made, bugs you already fixed, what happened last session, how the codebase fits together — kept in a file on your machine and handed back to the assistant next time, so you stop re-explaining your own project.
Works across Claude Code, Cursor, Windsurf, OpenAI Codex, OpenCode, Antigravity CLI, Aider, Goose, Cline, Roo Code, Crush, Pi and Hermes Agent — one memory, whichever tool you open.
No dependencies, no vector database, no background daemon. ~32MB of RAM, sub-millisecond lookups, works offline. Comparable tools install ~500MB of machine-learning libraries and take 200–500ms per lookup.
Why agi-memory? The 4 Cognitive Memory Pillars
Most AI memory architectures solve only a fragment of developer memory while incurring heavy dependencies or requiring background Node.js daemons. agi-memory unifies all four cognitive memory pillars in pure Python stdlib + SQLite (<35MB RAM, <1ms speed, zero external pip dependencies):
| Pillar | Core Question | Replaces | Implementation in agi-memory |
Latency / Overhead |
|---|---|---|---|---|
| 1. Epistemic | "What have we learned?" | Ad-hoc .cursorrules, forgotten bugfixes |
SessionLayer (SQLite FTS5 + BM25, Core Blocks) |
0.23 ms (zero tokens) |
| 2. Semantic | "What does our information mean & how is it connected?" | Heavy GraphRAG, Cognee, ChromaDB | GraphLayer (Native SQLite Recursive CTEs) |
0.28 ms (zero tokens) |
| 3. Episodic | "What happened during previous agent sessions?" | claude-mem (heavy Node/Bun daemons) |
EpisodicLayer (SQLite Session History & Lifecycle) |
0.23 ms (zero daemons) |
| 4. Structural | "How is this codebase structurally connected?" | Graphify, Tree-sitter binaries, LSP daemons |
CodeLayer (stdlib AST + Streaming Regex Graph) |
0.45 ms (zero daemons) |
Every pillar is scored by its own eval suite — see Benchmarks for measured comparisons against Mem0, Zep, Cognee, LangChain and claude-mem, including a real 13,988-observation production dataset.
Install
# One-line installer (recommended)
curl -fsSL https://raw.githubusercontent.com/kdbhalala/agi-memory/main/install.sh | bash
# Or: Homebrew / PyPI
brew tap kdbhalala/agi-memory https://github.com/kdbhalala/agi-memory && brew install agi-memory
pipx install agi-memory
Then wire up your assistants and initialize a project:
agi-integrate install all # configure every detected assistant + lifecycle hooks
agi-integrate status # confirm what was detected and configured
cd your-project && agi-integrate init .
init wires the project and installs an /agi-init slash command in each
assistant's own format. Run /agi-init inside your assistant and it reads the
codebase and writes the project's rules/ and context/ files.
Full options, including uvx and from-source: Installation.
Documentation
| Guide | What's in it |
|---|---|
| Installation | Installer script, Homebrew, PyPI/uvx, from-source, upgrading, hooks setup |
| The Four Pillars | Deep dive into L1 Epistemic, L2 Semantic, L3 Episodic, L4 Code Graph |
| Architecture | Layer boundaries, storage model, multi-assistant production layout |
| Supported Assistants | Per-tool config paths and rules files for all 13 assistants |
| CLI Usage | Every agi-memory and agi-integrate subcommand |
| Python API | Using the layers directly from Python |
| Vault & Git Sync | Append-only JSONL vault, cross-device sync, compaction |
| Benchmarks | Latency, memory and cost comparisons; real-dataset results |
| Testing & Evals | The L1-L4 eval suites, the actionability eval, chaos and stress tests |
| Integrations | Manual per-tool configuration snippets |
License
MIT
1 maintainer
Release files for agi-memory 0.6.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| agi_memory-0.6.1.tar.gz | 125.0 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| agi_memory-0.6.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 236.2 kB
Release files / agi_memory-0.6.1.tar.gz
| Download URL | agi_memory-0.6.1.tar.gz |
|---|---|
| Size | 125.0 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
3f32f7e70ba5afe5745532ebd5bf95de61b6c58bbe7b9258f9f9175ee1b616ae
|
|
BLAKE2b-256 checksum How to use checksums |
6aed3ff957102abe46adbe7fb9d624c41220a1c5f4c78a41f933cb381d16284f
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 14, 2026.
Transparency logRelease files / agi_memory-0.6.1-py3-none-any.whl
| Download URL | agi_memory-0.6.1-py3-none-any.whl |
|---|---|
| Size | 111.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
cca3de8bbe20c0cc976be6baf61328b24225e5f11920d9015227c4b3821b364e
|
|
BLAKE2b-256 checksum How to use checksums |
8af1d6ed15ab967263430fd4ed303290e9ad3579066a5aa7663e13f0187c48c1
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 14, 2026.
Transparency log
