Memoh Documentation
Memoh(/ˈmemoʊ/) is a multi-member, structured long-memory AI agent platform. Create bots with independent workspaces, persistent memory, tool access, and channel integrations across Telegram, Discord, Lark (Feishu), QQ, Matrix, Misskey, DingTalk, WeCom, WeChat, WeChat Official Account, Email, and Web.
Memoh ships in two forms:
- Desktop for personal and local use. The native app starts its own local server, embedded Qdrant, storage, and bundled CLI.
- Server Deploy for always-on, multi-user, multi-tenant, or channel-facing deployments.
Start Here
- Installation Overview — Choose Desktop or Server Deploy.
- Desktop Installation — Install the native local client.
- Server Deploy — Run Memoh with Docker Compose.
- Providers And Models — Set up providers, client types, and model roles.
- Bot Setup — Create a bot and configure its main tabs.
- Sessions — Understand chat vs discuss sessions, status panels, and routing.
Feature Guides
- Workspace Backends — Choose container and local workspace runtimes.
- Browser / Computer Use — Operate headed workspace browsers and GUI desktops.
- Files — Browse and edit the bot workspace filesystem.
- Channels — Platform overview and links to per-channel guides.
- Access Control — ACL presets, rule ordering, and scoped access.
- Skills — Managed skills, effective/shadowed states, and adoption workflow.
- Supermarket — Install skills and MCP templates from the catalog.
- MCP — Configure tool servers, OAuth, probing, import/export.
- Memory — Long-term memory providers and memory compaction.
- Context Compaction — Reduce active session context size without changing stored memory.
- Slash Commands — Learn the command model and daily control surface.