IHMT MEMORY

IHMT MEMORY provides long-term memory for AI coding agents as a tree of plain files on the local disk.

IHMT MEMORY is a universal, domain-agnostic long-term memory for large language models and AI coding agents. It stores decisions, environment setups, preferences, and corrections in a recursive tree of plain text and JSON files on the local disk. Instead of scanning a flat index, the agent walks the tree to retrieve only the relevant context, saving tokens and maintaining accuracy. When facts change, old entries are kept as history and flagged as outdated, while ambiguous queries prompt the user for clarification. The system operates locally without accounts, databases, or third-party cloud services. It supports agents such as Claude Code, Codex, and opencode through a standard Model Context Protocol server.

IHMT MEMORY at a glance

Best for: Software engineer
Runs on: macOS, Linux, API, self-hosted, open source
Pricing: Free
Released: October 2026
Category: AI coding tool
Website: github.com/gonzaroman/ihmt-memory

Key features

Stores long-term memory as a tree of plain text files
Shares memory across multiple AI agents and models
Tracks historical changes and flags outdated facts
Retrieves targeted context to save tokens
Runs locally without cloud storage or databases
Includes a graphical interface with a collapsible tree and timeline
Exposes tools for memory search, project indexes, and session scratch memory

Use it for

Remembering project decisions across multiple AI coding sessions
Sharing user preferences and environment setups between different AI agents
Tracking outdated facts and historical changes over time

Alternatives to IHMT MEMORY

AI Memory — Shared memory for coding agents that works across machines and team members.
IAI Memory Engine — Provides persistent, local, and encrypted memory for AI agents without APIs.
Tigerless Agent Memory — Maintains agent history in portable, greppable Markdown files without API keys.
LongMemory — Provides immutable, temporal, and governed recall instead of simple vector retrieval.
claude-mem — It shares persistent, searchable agent memory across sessions and devices.
Supermemory — Integrates long-term memory into AI coding agents with a single prompt.
Hy Memory — Provides specialized memory pipelines for chat, working, and coding tasks.
AgentMemory — It runs locally with zero external databases and achieves high retrieval recall.
Engram — Single Go binary with zero dependencies and built-in full-text search.
MemPalace — Achieves high recall with local-first storage and zero API calls.
Codebase Memory MCP — High-speed indexing and persistent knowledge graphs for AI coding agents.
memU — It gives every connected agent a shared and inspectable memory layer.

All IHMT MEMORY alternatives