mind: memory tool for Hermes Agent

Local, deterministic memory for coding agents, shipped as one dependency-free Python file

What mind does

The runtime is local, plain-text, offline and agent-neutral, built on the Python standard library and distributed as one auditable mind.py file. Facts can be semantic, episodic, procedural or decisions, and carry scope, authority, source trust, sensitivity, expiration and pinning. A journal tracks where each fact came from, unused salience ages over time, and a deterministic dream cycle consolidates recurring themes into cortex notes and writes human-readable receipts.

mind init writes a guard-marked standing-order block into AGENTS.md, CLAUDE.md and GEMINI.md and adopts existing Cursor, Windsurf, Cline and Roo rule files, preserving other content byte for byte. Agents then run python3 mind.py capture to store a durable fact. Automatic capture accepts stable decisions, conventions and environment facts, rejects credential and personal-identity patterns and transient task state, and quarantines untrusted material for review. The README lists 30 commands, a protocol server with 17 tools and stable release 7.0.0.

Key features

Weighted concept graph with typed facts and provenance journal
Recall through lexical ranking plus spreading activation
Deterministic dream cycle that consolidates recurring themes
Automatic capture that rejects credentials and transient task state
Standing orders exported to AGENTS.md, CLAUDE.md and GEMINI.md
Single mind.py file with 30 commands and a 17-tool protocol server

When to use it

Letting a coding agent remember project decisions and conventions between sessions
Keeping agent memory offline and auditable in plain text files
Sharing one memory store across several agents that read AGENTS.md

Who it is for: Developers who want persistent, inspectable project memory for coding agents without extra dependencies.

How it fits with Hermes Agent

Agent-neutral memory tool tagged hermes-agent; it works through project instruction files and a protocol server rather than a Hermes-specific plugin.

How to install mind

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

curl -fsSLO https://raw.githubusercontent.com/Da7-Tech/mind/v7.0.0/mind.py
python3 -c "import hashlib; p=open('mind.py','rb').read(); assert hashlib.sha256(p).hexdigest() == 'ae2fc389b3b09c93cb432ab55b71063d98b400da6b18d6bc178322bc8f3fcf69'"
python3 mind.py init

Requirements: Python 3 with only the standard library; on stock Windows the commands use py -3 mind.py

Note: The default branch is a 7.0.1.dev0 preview, and the README names 7.0.0 as the production-ready release.

FAQ

What is mind?

mind is a local memory system for coding agents. It stores durable project facts in a concept graph, recalls them on request and consolidates recurring themes through a deterministic dream cycle.

Does mind work with Hermes Agent?

It is agent-neutral, and the repository is tagged hermes-agent. The README describes standing orders written into AGENTS.md, CLAUDE.md and GEMINI.md plus a protocol server, but it does not describe a Hermes-specific setup.

How do I install mind?

Download mind.py from the v7.0.0 tag, verify its SHA-256 hash with the provided Python one-liner, then run python3 mind.py init. The README also shows a development install by cloning the repository.