immartian

immartian / bellamem

Public

Persistent belief-graph memory for AI agents. Retrieves decisive context by importance — not recency, not RAG, not /compact.

pypi.orgprojectbellamem agentic ai-agents belief-graph claude-code codex
10
0
100% credibility
Found Apr 13, 2026 at 10 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Bellamem provides persistent structured memory for AI agents by extracting beliefs, decisions, and causal chains from conversations into a queryable hypergraph that survives context resets.

How It Works

1
🔍 Discover Bella

You notice your AI helper forgets key decisions from yesterday's chat, like rejected fixes or project rules, and learn about Bella, a smart memory tool that remembers what matters.

2
📦 Get Started Easily

With one simple command, you add Bella to your project, and it quietly sets up a private memory space just for your work.

3
⚙️ Add Handy Shortcuts

You install quick commands that work right inside your AI chat, so Bella blends seamlessly into your daily workflow.

4
đź’ľ Save Your Sessions

Before ending a chat, you tell Bella to capture the important decisions, causes, and lessons learned, keeping them safe even after clearing your chat history.

5
🔄 Pick Up Where You Left Off

Next time, you ask Bella to remind you of what was decided, recent changes, and key facts, so your AI starts fresh but fully informed.

🧠 Your AI Remembers Forever

Now your helper recalls past fixes, avoids old mistakes, and builds on real progress across days, making work feel continuous and trustworthy.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 10 to 10 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is bellamem?

Bellamem is a Python library and CLI for persistent belief-graph memory in AI agents. It extracts structured beliefs—decisions, rejected ideas, causes, self-observations—from chat sessions, storing them in a hypergraph that survives /clear, /compact, and new contexts. Agents retrieve decisive context by importance and evidence mass, not recency or RAG, enabling continuity across days.

Why is it gaining traction?

It crushes benchmarks: 92% hit rate vs. 8% for /compact and 31% for RAG top-k on real coding queries. Claude Code slash commands like /bellamem save/resume/why make it dead simple—ingest sessions, resume with ~30k tokens of structured history, and an edit guard blocks re-suggesting disputes. Persistent storage beats ephemeral windows, scaling to 1800+ beliefs without collapse.

Who should use this?

Claude Code users debugging flaky tests or building agents that forget invariants. AI devs needing persistent memory for coding workflows, beyond basic github persistent storage hacks. Teams prototyping belief-graph agents where context must persist like empires, not vanish on reload.

Verdict

Alpha at 10 stars and 1.0% credibility—dogfooded on its own build, excellent docs and CLI, but light on tests. Grab it if you're chasing persistent memory github wins for agents; skip for production until v0.1 stabilizes.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.