2 months ago
Persistent memory for AI agents — captures session context automatically, compresses it (99% token savings with 14 content-aware summarizers), and retrieves relevant context in future sessions via hybrid search (BM25 + vector + optional LLM reranker).
Server Config
{
"mcpServers": {
"context-mem": {
"command": "npx",
"args": [
"-y",
"context-mem",
"serve"
]
}
}
}