Mnemoverse Memory — Claude Code MCP server: agent memory
A hosted long-term memory server for AI agents over MCP. Agents save notes and pull them back with a plain-language query.
claude mcp add --transport http mcp-memory-server https://mcp.mnemoverse.com/mcpWhat Mnemoverse Memory does
The server exposes memory_write to store something and memory_read to fetch it by description rather than exact match. A third call, memory_feedback, lets you mark a recalled note as helpful or misleading, and that verdict changes the ranking of later results. Passing domain: "user:<id>" on both read and write keeps separate users' notes apart; the field is a namespace inside your account, not an access control.
You can connect either way. The hosted endpoint https://mcp.mnemoverse.com/mcp authenticates over OAuth, so no key is involved. The local route runs npx -y @mnemoverse/mcp-memory-server with an API key supplied through environment variables. The free plan covers a thousand queries a day across reads, writes and feedback, with 10,000 atoms of storage and 60 requests per minute.
The memory engine itself is a hosted service. Only the MCP server is open, under MIT. Teams that must keep agent memory inside their own network won't get that here by default — running the engine yourself is available on Enterprise terms. Shared rooms, where several agents draw on one memory, are still in beta.
What it helps with
- keep an agent's decision and surface it in a later session
- separate each user's memories with the
domainfield - down-rank bad recalls through
memory_feedback - share one memory across different editors from a single account
- let several agents reuse a common pool in a shared room
Example prompts for the agent
- Save to memory that we picked PostgreSQL over SQLite and why
- Look up the commit message rules I gave you earlier
- Mark the last note about deployment as misleading
How to install Mnemoverse Memory
1Platform CLI: remote server
claude mcp add --transport http mcp-memory-server https://mcp.mnemoverse.com/mcp2Platform CLI
claude mcp add --env MNEMOVERSE_API_KEY=<MNEMOVERSE_API_KEY> --env MNEMOVERSE_API_URL=<MNEMOVERSE_API_URL> --transport stdio mnemoverse-memory -- npx -y @mnemoverse/mcp-memory-serverReplace <…> with your own values
1Platform CLI: remote server
codex mcp add mnemoverse-memory --url https://mcp.mnemoverse.com/mcp2Platform CLI
codex mcp add mnemoverse-memory --env MNEMOVERSE_API_KEY=<MNEMOVERSE_API_KEY> --env MNEMOVERSE_API_URL=<MNEMOVERSE_API_URL> -- npx -y @mnemoverse/mcp-memory-serverReplace <…> with your own values
Config file
{
"mcpServers": {
"mnemoverse-memory": {
"url": "https://mcp.mnemoverse.com/mcp"
}
}
}{
"mcpServers": {
"mnemoverse-memory": {
"command": "npx",
"args": [
"-y",
"@mnemoverse/mcp-memory-server"
],
"env": {
"MNEMOVERSE_API_KEY": "<MNEMOVERSE_API_KEY>",
"MNEMOVERSE_API_URL": "<MNEMOVERSE_API_URL>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"mnemoverse-memory": {
"type": "http",
"url": "https://mcp.mnemoverse.com/mcp"
}
}
}{
"servers": {
"mnemoverse-memory": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@mnemoverse/mcp-memory-server"
],
"env": {
"MNEMOVERSE_API_KEY": "<MNEMOVERSE_API_KEY>",
"MNEMOVERSE_API_URL": "<MNEMOVERSE_API_URL>"
}
}
}
}Replace <…> with your own values
1Platform CLI: remote server
gemini mcp add -t http mnemoverse-memory https://mcp.mnemoverse.com/mcp2Platform CLI
gemini mcp add -e MNEMOVERSE_API_KEY=<MNEMOVERSE_API_KEY> -e MNEMOVERSE_API_URL=<MNEMOVERSE_API_URL> mnemoverse-memory npx -y @mnemoverse/mcp-memory-serverReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Mnemoverse Memory
How do I install Mnemoverse Memory in Claude Code?
Run claude mcp add --transport http mnemoverse-memory https://mcp.mnemoverse.com/mcp, then start Claude and sign in through /mcp. To run the server locally instead, use the command with npx -y @mnemoverse/mcp-memory-server plus the MNEMOVERSE_API_KEY and MNEMOVERSE_API_URL variables.
What does Mnemoverse Memory do?
It keeps an agent's notes — decisions, preferences, lessons — and returns them when asked in ordinary language. Rating what came back shapes the next round of results.
Do I need an API key?
Not for the hosted endpoint: sign in with OAuth after creating an account. The key is for the local server that runs on your own machine.
