{
  "markdown": "# Mnema\n\n**Local-first, formally verified memory for AI agents — one shared store, every client, over [MCP](https://modelcontextprotocol.io).**\n\nYour agents rent their intelligence from a lab; their memory should be yours.\nMnema is a single, durable store of what your agents have learned — `consult` it\nbefore you act, `learn` a durable fact, walk the `related` graph, `list` what's\nrecent. Data lives in `~/.mnema` on your machine and nothing leaves it unless you\npoint it somewhere.\n\n- **Verified engine.** The embedded engram database is formally specified and\n  machine-checked (property catalog discharged by 250+ tests, an exhaustive CRDT\n  conformance model checker, Kani proofs, and a TLA⁺ model). Every released\n  binary passed that gate on its exact commit.\n- **Local-first.** An open, append-only segment format in `~/.mnema`. No cloud,\n  no account.\n- **One store, every client.** Claude Code, Claude Desktop, Cursor, your scripts,\n  and your CI all talk to the same memory. Recall is lexical + graph\n  spreading-activation (plus optional semantic embeddings), with an honest\n  confidence band on every answer.\n\n> This repository hosts the **public release bundles** (`.mcpb`) and prebuilt\n> binaries. The product, docs, and downloads live at **[mnema.zimac.ai](https://mnema.zimac.ai)**.\n\n## Install\n\n### From the MCP Registry (recommended)\n\nMnema is published to the official [MCP Registry](https://registry.modelcontextprotocol.io)\nas **`ai.zimac/mnema`**. Any registry-aware client can install it by name; it\npulls the platform `.mcpb` bundle from this repo's releases and verifies its\n`fileSha256`.\n\n### `.mcpb` bundle (manual)\n\nGrab the bundle for your platform from the [latest release](https://github.com/Zimac-AI/mnema/releases/latest)\nand open it with an MCP-bundle-aware client (e.g. Claude Desktop → Extensions):\n\n| Platform | Asset |\n|---|---|\n| macOS (universal) | `mnema-<version>-macos-universal.mcpb` |\n| Linux (x86_64) | `mnema-<version>-linux-x86_64.mcpb` |\n\nA `.mcpb` is a ZIP of a `manifest.json`, the `mnema` binary, and this README.\nEach is dry-run verified (the bundled binary completes an MCP `initialize`\nhandshake) before release.\n\n### Headless binary (`curl | sh`)\n\nFor agents, scripts, and CI — installs the standalone `mnema` binary:\n\n```sh\ncurl -fsSL https://mnema.zimac.ai/install.sh | sh\n```\n\n(macOS universal + Linux x86_64 prebuilt; anything else builds from source. The\nmacOS **GUI app** is a separate download at [mnema.zimac.ai](https://mnema.zimac.ai).)\n\n## Give your agents memory\n\n```sh\nclaude mcp add mnema -- mnema mcp\n```\n\nThat's the whole setup: `mnema mcp` auto-starts the shared local store if it\nisn't running, and the session gets four tools — `consult` (search before you\nact), `learn` (record a durable fact), `related` (graph neighbours), `list`\n(browse). Any MCP client works; the generic config is:\n\n```json\n{ \"mcpServers\": { \"mnema\": { \"command\": \"mnema\", \"args\": [\"mcp\"] } } }\n```\n\n### Point it at a shared / team store\n\nThe same binary bridges to a remote Mnema (a token-protected `mnema serve` or a\nhosted learnings pod) instead of the local one:\n\n```sh\nMNEMA_URL=https://memory.example.com MNEMA_TOKEN=<secret> mnema mcp\n```\n\nWith a non-loopback URL nothing is auto-started — it's a pure authenticated\nbridge.\n\n## Configuration\n\n| Env | Default | What |\n|---|---|---|\n| `MNEMA_DATA_DIR` | `~/.mnema` | Where the store lives |\n| `MNEMA_URL` | `http://127.0.0.1:7077` | Store the CLI/MCP bridge targets |\n| `MNEMA_TOKEN` | — | Bearer credential for a remote store |\n\n## Links\n\n- **Product & downloads:** https://mnema.zimac.ai\n- **Registry entry:** `ai.zimac/mnema` on https://registry.modelcontextprotocol.io\n- **Releases (`.mcpb` + binaries):** https://github.com/Zimac-AI/mnema/releases\n\n---\n\nMnema is a [Zimac](https://zimac.ai) project.\n",
  "bytes": 3812,
  "sha": "8a3f89524c01a37116682ff483ef28e5a9bdafecc20e227b99e221491be6667d",
  "repo_slug": "zimac-ai/mnema",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_ai_zimac_mnema_1b66be40/readme"
}