supermemoryai/opencode-supermemory

What it solves

It provides a persistent memory layer for AI coding agents (specifically for OpenCode), allowing them to remember user preferences, project-specific knowledge, and previous conversations across different sessions and projects. This prevents the agent from forgetting critical project details or user requirements every time a new session starts.

How it works

The plugin integrates with the Supermemory service to store and retrieve information. It uses several mechanisms to maintain context:

• Context Injection: Automatically provides the agent with a user profile and relevant project memories at the start of a session.
• Reasoned Recall: The agent silently decides on every turn whether it needs to search its memory to answer the current prompt.
• Automatic Capture: Periodically saves conversation turns and redacts private information.
• Keyword Detection: Triggers memory saves when the user uses phrases like "remember this".
• Codebase Indexing: A specialized command (/supermemory-init) allows the agent to explore and memorize the project structure.
• Preemptive Compaction: When the context window is nearly full, it summarizes the session and saves that summary as a memory to preserve continuity.

Who it’s for

Developers using OpenCode who want their AI agent to have long-term memory of their coding standards, project architecture, and past interactions.

Highlights

• Cross-Session Persistence: Memories are maintained across different projects and sessions.
• Semantic Search: Uses a supermemory tool for the agent to search and recall information.
• Privacy Controls: Content wrapped in <private> tags is never stored.
• Flexible Scoping: Distinguishes between personal user memories and project-specific memories.
• Interoperability: Compatible with memory containers used by other tools like Claude Code and Codex.