Memanto MCP exposes typed persistent memory to assistants through a Python bridge with a Moorcheh API key and an explicit agent namespace.
Memanto MCP exposes typed persistent memory to assistants through a Python bridge with a Moorcheh API key and an explicit agent namespace.
The separate memanto-mcp package exposes memory storage, retrieval and grounded-answer tools. A configured default agent identifier lets several clients work with the same namespace. This bridge is distinct from the broader Memanto CLI and migration workflows in the repository.
The standard client launch uses memanto-mcp over stdio with MOORCHEH_API_KEY. MEMANTO_DEFAULT_AGENT_ID selects the namespace used when a call omits an agent identifier. The package also supports optional network transports. Its API key authenticates upstream access; a network listener needs its own inbound access controls rather than assuming the upstream key protects every connecting client.
Store a fictional preference, retrieve it from another session and inspect the source and provenance fields. A corrected memory should retain enough context to distinguish an explicit statement from an inference. Agent-management tools are optional and are not enabled by the default seven-tool configuration. Keep administration access separate from routine remember and recall workflows.
Check the linked publisher documentation and your connected client's tool list. Names and availability can vary with the server version and account permissions.
| Tool | What it does |
|---|---|
| remember | Store a typed memory with provenance. |
| batch_remember | Store a batch of memories. |
| recall | Search the selected memory namespace. |
| recall_recent | Retrieve recent memories. |
| recall_as_of | Retrieve memory state for a selected point in history. |
| recall_changed_since | Retrieve memory changes after a reference point. |
| answer | Generate a grounded answer from retrieved memory. |
{
"mcpServers": {
"memanto": {
"command": "memanto-mcp",
"args": [],
"env": {
"MOORCHEH_API_KEY": "YOUR_MOORCHEH_API_KEY",
"MEMANTO_DEFAULT_AGENT_ID": "test-assistant"
}
}
}
}Python 3.10 or later, memanto-mcp with its compatible dependencies and a Moorcheh API key.
| Variable | Description | Required |
|---|---|---|
| MOORCHEH_API_KEY | Moorcheh API key for the configured memory service. | Yes |
| MEMANTO_DEFAULT_AGENT_ID | Recommended default memory namespace. | Optional |
Store entities, relations and observations in a local knowledge graph with the reference Memory MCP server.
A structured scratchpad for hard problems — stepwise reasoning with revisions, branches and visible logic.
Fourteen research tools, from arXiv to SEC filings. Tool calls are billed to your Apify account.
Vault-native, accountable memory for Claude Code and MCP clients. Markdown is ground truth.
CCXT integration for Model Context Protocol
bioRxiv/medRxiv/arXiv preprint full text as structured sections for AI agents + search.