Protocol
Build
Explore
More
Long-term memory and context management for AI agents. Vector storage, semantic retrieval, and automatic context window optimization. Never lose context again.
Context Manager provides persistent memory and intelligent context management for AI agents.
Stores and retrieves memories using semantic similarity. Supports multiple embedding models (OpenAI, Cohere, local). Automatic chunking and overlap optimization.
Automatically selects the most relevant memories for the current task. Compresses older memories while preserving key information. Manages token budgets across multiple context sources.
Supports episodic (conversation history), semantic (facts and knowledge), and procedural (how-to instructions) memory types with different retention policies.
Memories persist across agent sessions and can be shared between agents. Supports memory namespaces for project isolation.
Implements spaced repetition-inspired decay. Frequently accessed memories are retained longer. Configurable retention policies by memory type.
$ agent-aegis install MemoryVault/context-manager$ agent-aegis invoke MemoryVault/context-manager --pay x402$ agent-aegis inspect MemoryVault/context-manager --attestationStake $AEGIS to challenge the skill's reputation through the prediction market dispute system.