smarzola

smarzola / hermes-local-memory

Public

Local-first SQLite memory provider for Hermes Agent

agent-memory ai-agents hermes-agent local-first sqlite
14
0
100% credibility
Found Apr 26, 2026 at 14 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Hermes Local Memory is a local-first SQLite memory system for the Hermes Agent that stores conversation history, facts, profiles, and cards in an inspectable database controlled by the agent.

How It Works

1
📰 Find local memory helper

You discover a simple tool that gives your AI assistant a private memory stored right on your computer, no servers needed.

2
📦 Install the tool

You grab the tool with a quick install command, and it's ready on your machine.

3
🔗 Connect to your AI

You link the memory tool to your Hermes assistant setup, so it starts using local storage seamlessly.

4
📥 Bring in old memories

You safely copy over past chats and facts from other systems, previewing everything first to check it looks right.

5
🔍 Peek inside

You browse profiles, facts, cards, and search memories using friendly list commands, seeing exactly what's saved.

6
🧹 Tidy and improve

You review and organize memories, letting the AI suggest cleanups or adding new insights with safe previews.

🎉 AI remembers perfectly

Your assistant now recalls people, preferences, and conversations reliably from your local memory, feeling smarter and more personal.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 14 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 hermes-local-memory?

Hermes Local Memory is a Python SQLite provider for Hermes Agent, delivering local-first memory storage in a single `~/.hermes/memory/local_memory.sqlite` file. It handles peer profiles, raw history, durable facts, compact cards, search, and context injection—letting agents maintain memory via explicit tools like `memory_get_card` or `memory_conclude`. Users get auditable, agent-controlled persistence without servers, queues, or opaque "dream" backends.

Why is it gaining traction?

It ditches magical vector stores and daemons for boring SQLite local-first sync, with CLI commands like `hermes-local-memory peers --json` for inspection and `import honcho-api` for safe migrations. Deterministic consolidation, peer review packets, and reflection tools keep facts conservative and source-labeled, appealing to devs who want transparency over hidden mutations. The Hermes plugin shim installs in one command, hooking seamlessly into agent workflows.

Who should use this?

Hermes Agent runners seeking local-first GitHub alternatives to server-based memory like Honcho. Perfect for solo devs or teams building agent memory providers with sqlite local first sync, especially during migrations—dry-run imports and CLI repairs handle noisy data without production risks. Avoid if you need distributed scale.

Verdict

Alpha with 13 stars and 1.0% credibility score, but strong docs, tests, and CLI lower the trial barrier—import to a test DB first. Worth adopting for local, inspectable Hermes memory if servers feel like overkill.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.