claude-mem

Persistent, cross-agent memory for Claude Code and other coding assistants

Freemium· claude-mem (open-source): Free (Apache-2.0) / CMEM Cloud: $20 per month / Team: $333 per seat/month (3-50 seats, includes Forward Deployed Engineer support)AgentsModel-agnostic — works with the models used by the connected agent (Claude, GPT, Gemini, etc.) via MCP

Visit website →

Best for

Solo developers and small engineering teams heavily using Claude Code, Cursor, or other MCP-based coding agents who want persistent, searchable context that survives across sessions, machines, and even different agent vendors.

Skip if

Non-technical users, teams that do not use MCP-compatible agents, or organisations with strict data-residency rules that cannot self-host the open-source engine instead of using CMEM Cloud.

claude-mem (paired with the managed CMEM Cloud service at cmem.ai) is an agent memory layer that captures the observations, decisions, and dead ends produced during coding-agent sessions and makes them queryable later, across machines, IDEs, and different agents. The open-source engine runs locally, writes structured memories to a local vector store, and exposes them to any Model Context Protocol (MCP) client; the optional cloud service syncs those memories through a private MCP link so context follows the developer between Claude Code, Cursor, ChatGPT, Gemini, and CLI agents without manual re-briefing. It is aimed at individual engineers who want their agent to remember prior architecture choices and gotchas, and at small teams who want shared, scoped memory plus routing of relevant context ('brainbeats') to the right team member. Typical workflows include: bootstrapping a new session on a repo with the previous run's decisions already loaded, sharing a rediscovered bug workaround with a teammate's next agent invocation, running semantic search over months of temporal notes, and viewing a live mobile dashboard of what agents are learning. It ships with an offline-first architecture, eleven bundled skills, per-project/per-member scoping, and Stripe-based checkout for the paid tiers. The open-source component is on GitHub under Apache-2.0, so self-hosting the memory engine without ever touching the cloud is a supported path.

Editor's take

This is one of the more thoughtful takes on agent memory I have seen: an open-source local engine plus an optional cloud sync, rather than a walled SaaS. If you already run Claude Code or Cursor daily, the $20/month tier is easy to justify; if you don't, the free self-hosted path is still worth wiring in. The team tier's pricing is aggressive and probably only makes sense once shared context is genuinely a bottleneck.

— The AI Tool Bible editorial team

Pros

✅ Open-source core (Apache-2.0) means the memory engine can be self-hosted and audited
✅ Cross-agent by design via MCP — memories carry across Claude Code, Cursor, ChatGPT, Gemini, and other MCP clients
✅ Offline-first with a local vector store, so nothing has to leave the machine unless cloud sync is enabled
✅ Structured capture of observations, decisions, and dead ends is more useful than raw transcript logging
✅ Scoped memory per project and per team member avoids the usual 'one giant global memory' contamination
✅ Includes 11 bundled skills and a live mobile dashboard out of the box

Cons

⚠️ Value depends on the user already living inside MCP-capable coding agents — casual chat users get little from it
⚠️ Team plan at $333/seat/month is a steep jump from the $20 individual tier
⚠️ Branding is split between claude-mem, CMEM, and cmem.ai, which makes docs and pricing harder to navigate
⚠️ Memory quality is only as good as what the agent chooses to record — noisy sessions can pollute future context
⚠️ Cloud sync introduces a trust boundary that some enterprises will need to review before adopting

Use cases

Persistent memory for Claude Code sessionsCross-IDE context sharing via MCPSemantic search over past agent decisionsTeam knowledge routing between engineersRecording architectural decisions and dead endsSelf-hosted local memory store for coding agentsMobile dashboard for agent activityScoped per-project agent memory