engineering / skill

Claude-Mem

Keep context across coding sessions by capturing tool usage, summarizing it and making it available to future sessions.

At a glance

Purpose
Keep context across coding sessions by capturing tool usage, summarizing it and making it available to future sessions.
Runs in
Chat app with skills, Local agent with shell, MCP server, Web service
Operating systems
Not stated
Needs
Runtime: Node.js 20.0.0+, Bun, uv and SQLite 3; Needs internet: Not stated; Needs accounts or API keys beyond your agent: Not stated
Files and data
Reads files: Yes; Writes files: Yes; Sends data to other services: Yes
Tool type
Skill
Delivery
From GitHub
Version
Latest on the main branch
Access
Code on GitHub
Compatibility
According to the README, it works with Claude Code and offers installs for OpenCode (1.3.4 or later, including OpenCode 2), T3 Code with Codex and Claude Code providers, Antigravity CLI, OMP (Oh My Pi), Pi, DeepSeek Harness, Grok Bot and OpenClaw gateways. The GitHub description also names Codex, Gemini, Hermes and Copilot.
External costs
The repository is Apache-2.0-licensed. According to the README, the default install offers a 30 day free trial of the claude-mem observer, after which memory falls back to your Anthropic plan unless you subscribe. You can instead use your own OpenRouter or Gemini key, or your Anthropic plan. The README also describes a cloud sync, and says the Grok Bot install defaults to CMEM Pro, a hosted memory service, with a local observer as an opt-in; it states no prices. Your coding agent or model provider may charge for usage.
Permissions
The software acts through your coding agent with the access you give it. According to the README, it installs lifecycle hooks that capture tool usage, runs a local worker service with an HTTP API and a web viewer managed by Bun, stores data in SQLite and a Chroma vector database, and keeps settings in ~/.claude-mem/settings.json. The README says Bun and uv are installed automatically if missing. For Grok Bot it says the plugin watches chat log files and appends some observations to that bot's memory log, which you can disable. The OpenClaw route is a one-line installer that pipes a downloaded script to bash. The README says private tags exclude content from storage.
Usage terms
Apache License 2.0, as stated in the repository. See its LICENSE file for the exact terms. The README also points to docs/license.md and docs/ip-boundary.md for licensing scope and what it calls the open/commercial boundary.

Code on GitHub

thedotmack/claude-mem

Apache-2.0

TypeScript100,113 stars8,783 forksLast pushed 13 Oct 2026 UTC

Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More

Topics: ai ai-agents ai-memory anthropic artificial-intelligence chromadb claude claude-agent-sdk

Public GitHub data fetched · 0 days old

View on GitHub ↗

What it does

Claude-Mem is a persistent memory system for Claude Code. According to the README, it automatically captures tool usage observations, generates semantic summaries and makes them available to future sessions, so the agent keeps knowledge of a project after a session ends or reconnects.

The README lists these parts: five lifecycle hooks, a local worker service with a web viewer and search endpoints, a SQLite database for sessions, observations and summaries, a mem-search skill for natural language queries, and a Chroma vector database for hybrid semantic and keyword search. It describes MCP search tools used in three steps: search returns a compact index, timeline gives context around a result, and get_observations fetches full details.

The repository also holds installers and plugins for other hosts. The skill folders sit under plugin, claude-mem-cursor, claude-mem-grok-bot, cowork and openclaw, and include mem-search, make-plan, do, handoff, standup and smart-explore.

Who it is for

Developers who use Claude Code or another supported coding agent and want it to remember earlier sessions.

Example tasks

Install Claude-Mem so context from earlier sessions appears in new sessions.
Search your project history with the mem-search skill.
Fetch full details for chosen observations with get_observations.
Use private tags, which the README says exclude sensitive content from storage.
Set CLAUDE_MEM_MODE to change the workflow mode or the language of generated observations.

Getting started

According to the README, run npx claude-mem install and restart Claude Code. The README gives host-specific forms such as npx claude-mem install --ide opencode, npx claude-mem install --ide t3code, npx claude-mem install --ide antigravity and npx claude-mem install --ide grok-bot. Inside Claude Code you can instead run /plugin marketplace add thedotmack/claude-mem and then /plugin install claude-mem. For OpenClaw the README gives a one-line shell installer. The README has the exact commands.

Limitations

According to the README, Claude-Mem needs Node.js 20.0.0 or higher and the latest Claude Code version with plugin support. Bun and uv are auto-installed if missing, and SQLite is bundled. The README says the default install asks you to sign in to a claude-mem account in a browser by email magic link, which starts a 30 day free trial of the claude-mem observer. When the trial ends, memory falls back to your Anthropic plan unless you subscribe. You can skip sign-in with an explicit provider flag, by setting CLAUDE_MEM_ONLINE_OPTIN to false, or in CI or non-interactive shells. The README says npm install -g claude-mem installs only the SDK and library, not the hooks or worker service. Restart Claude Code after installing. The repository can change at any time; check its README before you install.

Data practices

According to the README, the worker service and SQLite database store sessions, observations and summaries, and private tags exclude content from storage. After sign-in you choose the memory provider: the claude-mem observer, your own OpenRouter or Gemini key, or your Anthropic plan. The README describes a cloud sync that backs up your memories online, with the worker syncing on write. It says the OpenClaw installer can set up optional real-time observation feeds to Telegram, Discord, Slack and more, and that the Grok Bot install defaults to the hosted CMEM Pro memory. The README does not describe what each provider receives or keeps, and it does not mention telemetry.

Community experience

Not rated yet

Sign in to share your experience.

Comments 0

No comments yet.

Sign in to comment

Report a listing concern