RELATED 01
Engram
Gives coding agents a persistent, searchable project memory stored locally in SQLite
7.8kGitHub stars
5.7k/wknpm downloads
87popularity score
agentmemory
Gives coding agents a persistent memory they can save to, search and share across sessions
By Rohit Ghumare (individual) JavaScript Local · stdio Knowledge and memoryDeveloper tools
78 popularity score ★ 30k ↓ 19k/wk
Connect your client
Local (npm)
Add in one click
Opens the app, which asks before it adds the server.
Or add it by command or config
Claude CodeTerminal
claude mcp add agentmemory -- npx -y @agentmemory/mcpClaude Desktopclaude_desktop_config.json
{
"mcpServers": {
"agentmemory": {
"command": "npx",
"args": [
"-y",
"@agentmemory/mcp"
]
}
}
}Cursor~/.cursor/mcp.json
{
"mcpServers": {
"agentmemory": {
"command": "npx",
"args": [
"-y",
"@agentmemory/mcp"
]
}
}
}VS Code.vscode/mcp.json
{
"servers": {
"agentmemory": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@agentmemory/mcp"
]
}
}
}Codex~/.codex/config.toml
[mcp_servers.agentmemory]
command = "npx"
args = ["-y", "@agentmemory/mcp"]Gemini CLITerminal
gemini mcp add --scope user agentmemory npx -- -y @agentmemory/mcpSettings 2
AGENTMEMORY_URL |
Optional: Address of a running agentmemory server (e.g. http://localhost:3111); without it the shim exposes only 7 local tools. |
|---|---|
AGENTMEMORY_FORCE_PROXY |
Optional: Set to 1 for sandboxed clients that cannot reach the host's localhost, together with a reachable AGENTMEMORY_URL. |
What it does
agentmemory is a memory server for coding agents such as Claude Code, Cursor, Codex CLI, Gemini CLI, OpenCode and any other MCP client. It records observations from agent sessions and lets the assistant save insights, decisions and patterns, then recall them later with keyword (BM25), semantic and graph search. Every connected agent shares the same memory store.
Beyond recall it keeps a timeline of sessions, project profiles, a relationship graph, links between sessions and git commits, an audit trail, snapshots, and tools for multi-agent work: action items with dependencies, leases, signals between agents and peer-to-peer sync between instances. The full server exposes 54 tools, 6 resources and 3 prompts.
The npm package @agentmemory/mcp is a thin shim over @agentmemory/agentmemory. It gives the full tool set only when it can reach a running agentmemory server through AGENTMEMORY_URL (by default http://localhost:3111); with no server reachable it falls back to a local set of 7 tools. Starting the full server needs Node.js 20 or newer and runs a pinned iii-engine. It works with no API key (keyword search only); local embeddings, or an LLM or embedding provider key such as Anthropic, OpenAI, Gemini or Voyage, add semantic search and compression.
What it needs
| Runtime | Node.js |
|---|
Tools 35
memory_recall- Search past observations
memory_save- Save an insight, decision, or pattern
memory_smart_search- Hybrid semantic + keyword search
memory_sessions- List recent sessions
memory_compress_file- Compress markdown files while preserving structure
memory_file_history- Past observations about specific files
memory_patterns- Detect recurring patterns
memory_vision_search- Search image observations
memory_timeline- Chronological observations
memory_profile- Project profile (concepts, files, patterns)
memory_export- Export all memory data
memory_relations- Query relationship graph
memory_commit_lookup- Sessions behind a git commit
memory_commits- Commits recorded for a session
memory_graph_query- Knowledge graph traversal
memory_consolidate- Run 4-tier consolidation
memory_claude_bridge_sync- Read MEMORY.md or create a project export, preserving existing files
memory_team_share- Share with team members
memory_team_feed- Recent shared items
memory_audit- Audit trail of operations
memory_governance_delete- Delete with audit trail
memory_snapshot_create- Git-versioned snapshot
memory_action_create- Create work items with dependencies
memory_action_update- Update action status
memory_frontier- Unblocked actions ranked by priority
memory_next- Single most important next action
memory_lease- Exclusive action leases (multi-agent)
memory_signal_send- Inter-agent messaging
memory_signal_read- Read messages with receipts
memory_mesh_sync- P2P sync between instances
memory_diagnose- Health checks
memory_heal- Auto-fix stuck state
memory_facet_tag- Dimension:value tags
memory_facet_query- Query by facet tags
memory_verify- Trace provenance