Cairn

About

Cairn acts as a unified, persistent memory system designed for both AI agents and humans, preventing the loss of valuable insights and context. It automatically captures agent decisions, learnings, and session progress, enabling continuity across multiple interactions. Simultaneously, it offers intuitive tools for humans to quickly capture thoughts, URLs, and notes, centralizing all knowledge into a single, searchable pool. The system leverages AI for smart ingestion, auto-classification, enrichment, and advanced hybrid search, ensuring that every piece of information is stored once and always discoverable.

Key Features

• 8 GitHub stars
• Session continuity for AI agents via 'cairns' and narrated event timelines.
• Quick capture for humans with slash commands, URL extraction, and mobile/browser integrations.
• Smart ingestion pipeline for text and URLs, including auto-classification, chunking, and deduplication.
• Hybrid search combining vector similarity, full-text, and tag matching with Reciprocal Rank Fusion.
• AI-powered auto-enrichment, pattern discovery (HDBSCAN clustering), and relationship linking for memories.

Use Cases

• Enable AI agents to remember past decisions, learnings, and session context across multiple interactions.
• Provide a unified system for humans to quickly capture, organize, and retrieve personal or project-related knowledge.
• Automate the enrichment and contextualization of captured information using large language models.