mcp-memory-service-lite is an open-source semantic memory service designed for AI agents. It uses ONNX embeddings for efficient storage and retrieval, offers a REST API and MCP transport, and requires no PyTorch. Suitable for developers needing lightweight, cross-platform AI memory solutions.
mcp-memory-service-lite is a Memory & skills project. It focuses on providing a lightweight, efficient semantic memory service for AI agents without heavy dependencies. It is built as an open-source project for AI developers building agent memory systems. mcp-memory-service-lite is open source under the Apache-2.0 license. mcp-memory-service-lite is available on the command line and API.
doobidoo builds and maintains mcp-memory-service-lite, and it first shipped in 2026. Among its 5 catalogued features are semantic memory, ONNX embeddings, and REST API. It exposes integrations via an MCP server and a public API.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do