# 0Latency Memory

Persistent memory API for AI agents. Store, search, and recall memories with sub-100ms retrieval on cached queries. Works with Claude Desktop, Claude Code, Claude.ai, Cursor, Windsurf, and any MCP-co…

## Quick Start

```bash
# Connect this server (installs CLI if needed)
npx -y @smithery/cli@latest mcp add zerolatency/memory

# Browse available tools
npx -y @smithery/cli@latest tool list zerolatency/memory

# Get full schema for a tool
npx -y @smithery/cli@latest tool get zerolatency/memory memory_add

# Call a tool
npx -y @smithery/cli@latest tool call zerolatency/memory memory_add '{}'
```

## Direct MCP Connection

Endpoint: `https://memory--zerolatency.run.tools`

**Required config:**
- `Key` (query) — Persistent memory API for AI agents.

## Tools (7)

- `memory_add` — Extract and store memories from a conversation turn. Provide the human message, agent response, and an agent_id to name…
- `memory_recall` — Recall relevant memories given a conversation context. Returns a formatted context block ready to inject into a prompt.
- `memory_search` — Search memories by text query. Returns matching memories ranked by relevance.
- `memory_list` — List stored memories with optional filters. Supports pagination via limit/offset.
- `memory_delete` — Delete a specific memory by its ID.
- `memory_graph` — Query the knowledge graph. List entities, explore an entity's relationships, or find the path between two entities.
- `list_agents` — List all agent namespaces for this tenant with memory counts. Useful for discovering available agents.

```bash
# Get full input/output schema for a tool
npx -y @smithery/cli@latest tool get zerolatency/memory <tool-name>
```