-
Notifications
You must be signed in to change notification settings - Fork 0
Home
Zhafron Kautsar edited this page Jan 12, 2026
·
5 revisions
Welcome to the OpenCode Memory documentation. This wiki provides comprehensive guides for installation, configuration, and usage of the persistent memory system for AI coding agents.
Major Changes:
- User-Scoped Memories Removed: All memories are now project-scoped by default
-
Simplified Configuration: Consolidated
maxMemories(removedmaxProjectMemories) -
Renamed Settings:
userMemoryAnalysisInterval→userProfileAnalysisInterval -
Cleaner API: Removed
scopeparameter from all memory operations - Streamlined UI: Simplified stats and forms (no more user/project distinction)
Breaking Changes: See Configuration Guide for migration instructions.
Major Changes:
- Structured User Profile System: Replaced user-scoped memories with comprehensive profile system
- Profile Components: Preferences (with confidence), patterns (with frequency), workflows, skill assessment
- Profile Versioning: Changelog tracking with snapshot support
- Confidence Decay: Automatic preference confidence decay over time
- Web UI Profile Viewer: New interface for viewing profile and changelog
Major Changes:
- Prompt-Based Auto-Capture: Replaced token-based system with simpler prompt-based approach
- Unified Timeline: Browse memories and prompts together with visual linking
- Prompt-Memory Linking: Bidirectional links between prompts and generated memories
- User Learning System: Analyzes user patterns from conversation history
- Multi-Provider AI: Support for OpenAI, Anthropic, and OpenAI-compatible APIs
- Installation Guide - Step-by-step installation instructions
- Quick Start - Get up and running in 5 minutes
- Configuration Guide - Complete configuration reference
- User Profile System - Automatic learning system
- Memory Operations - Using the memory tool
- Web Interface - Managing memories through the UI
- Auto-Capture System - Automatic memory extraction
- API Reference - REST API and tool reference
- Embedding Models - Local and API-based models
- Performance Tuning - Optimization strategies
- Troubleshooting - Common issues and solutions
OpenCode Memory is a plugin that provides AI coding agents with persistent memory across sessions. It uses vector embeddings and SQLite to store and retrieve contextual information efficiently.
- Project Memory System: Persistent storage for project-specific knowledge
- Structured User Profile: Automatic learning of preferences, patterns, workflows, and skills
- Vector Search: Fast similarity-based retrieval
- Unified Timeline: Browse memories and prompts together
- Prompt Linking: Track memory origins with bidirectional links
- Web Interface: Full-featured UI for memory management
- Auto-Capture: Intelligent prompt-based memory extraction
- Local-First: All data stored on your machine
- Zero Config: Works out of the box with sensible defaults
- Bun runtime (recommended) or Node.js 18+
- OpenCode AI platform
- 2GB RAM minimum (4GB recommended)
- 1GB disk space for database
If you encounter issues or have questions:
- Check the Troubleshooting guide
- Search existing GitHub Issues
- Open a new issue with detailed information
Contributions are welcome. Please see the main README for contribution guidelines.