OpenCode Memory
An OpenCode plugin that gives coding agents persistent, searchable memory across sessions, backed by a local vector database with no separate server to run.
OpenCode Memory is a persistent memory system for AI coding agents built as an OpenCode plugin. It stores project memories, an automatically learned user profile, and a combined memory-and-prompt timeline in an embedded Turso and libSQL database with native vector indexes, so no separate vector database or custom SQLite build is required.
Memory is searched with an approximate nearest-neighbor vector index, and features that require structured output, such as automatic capture and profile learning, need an AI provider that supports tool calls; plain memory search, add and list operations work without one. The plugin includes a web UI for browsing project memory and user profiles, supports OpenAI and Anthropic as providers, offers more than a dozen local embedding models, and migrates older SQLite-shard memory stores to the native vector format automatically. It is written in TypeScript and licensed under MIT.