Discover and explore top open-source AI tools and projects—updated daily.
supermemoryaiAI agent with real-time learning and persistent memory
Top 20.3% on SourcePulse
This plugin enables Claude Code to gain persistent, real-time learning capabilities by integrating with Supermemory. It allows AI agents to retain knowledge across sessions and projects, enhancing user productivity and AI utility by remembering past work and project context.
How It Works
The plugin leverages Supermemory Pro to provide state-of-the-art memory for Claude Code. On session initiation, it fetches and injects relevant memories, such as user profiles and recent project context, into Claude's active session. During conversations, turns are automatically captured and stored. Additionally, it offers codebase indexing to understand project architecture, patterns, and conventions, creating a continuously evolving knowledge base for the AI.
Quick Start & Requirements
Installation involves adding the plugin via /plugin marketplace add supermemoryai/claude-supermemory or from a local path, followed by /plugin install claude-supermemory. A Supermemory Pro account and an API key (SUPERMEMORY_CC_API_KEY) obtained from console.supermemory.ai are mandatory. Optional environment variables like SUPERMEMORY_SKIP_TOOLS and SUPERMEMORY_DEBUG allow customization, alongside a ~/.supermemory-claude/settings.json file for fine-grained control over captured tools and profile items.
Highlighted Details
/claude-supermemory:index command.super-search Skill: Enables agents to recall past work and information by querying their stored memories.Maintenance & Community
No specific details regarding maintenance, community channels, or notable contributors are provided in the README.
Licensing & Compatibility
The project is released under the MIT license, which generally permits broad use, including commercial applications and integration into closed-source projects.
Limitations & Caveats
A primary requirement is a Supermemory Pro subscription. Users can configure which tools' outputs are captured or skipped, potentially limiting the scope of learned context if not set appropriately.
1 week ago
Inactive