Memory for AI agents
Animus
Animus — AI Memory Synchronization Network connects different AI assistants, collects and indexes context produced across machines and turns validated knowledge into a curated Obsidian base. The self-hosted hub offers search, peer management, automatic updates and IDE integration through Model Context Protocol.
Explore features and technologies
Secure peer synchronization
Each machine operates as an autonomous peer and synchronizes context with a self-hosted hub over HTTPS, with signed requests and replay protection.
Instant full-text search
A SQLite FTS5 index across thousands of chat turns finds decisions, snippets and patterns in milliseconds, even across workspaces.
Knowledge OS in Obsidian
Validated knowledge is organized in a curated Obsidian base, separating raw memory, confirmed decisions and reusable documentation.
Unified multi-AI memory
Unifies memories from GitHub Copilot Chat and CLI, Claude Code and Antigravity — one assistant can reuse context produced by the others.
Peer auto-update and repair
Peers automatically check releases and can receive an idempotent ZIP package prepared by the hub for installation, updates or repairs.
Contextual RAG through MCP
The MCP interface provides search and context to IDE agents; on the web, Markdown documents get secure rendering with Mermaid diagrams.