LocalRecall is an open-source memory layer and knowledge base for AI agents that runs entirely on your own hardware, with 981 stars on GitHub and active development through 2026.
LocalRecall is a lightweight, no-frills RESTful API for managing knowledge bases and files stored in vector databases โ no GPU, no internet connection and no cloud services required. It provides a simple, generic abstraction layer over knowledge retrieval, so AI agents and chatbots can manage both long-term and short-term memory without depending on a hosted provider. It comes from the same author as LocalAI and is designed to slot into local-first AI stacks where data privacy and offline operation matter. Because everything runs locally, your documents and embeddings never leave your machine unless you choose to connect it to something that does. It stores embeddings alongside the original files, and because the abstraction layer is generic, the same agent memory code can move from a laptop-grade file store to a production PostgreSQL deployment without changing how the agent talks to it.
LocalRecall is completely free and open source under the MIT license. You download it, run it on your own hardware and pay nothing โ there is no cloud tier, no usage metering and no account required.
LocalRecall suits developers building agents or chatbots that need persistent memory without sending user data to a third-party service. It is a strong fit for privacy-sensitive projects, offline deployments and anyone already running a local AI stack who wants a simple, self-hosted retrieval layer.