Indexes and searches local coding agent history across multiple platforms to provide a unified knowledge base for AI-assisted development.
CASS (Coding Agent Session Search) is a high-performance utility designed to centralize and search historical data from various AI coding agents like Claude Code, Cursor, Aider, and Codex. It provides a unified CLI/TUI that allows both human developers and AI agents to retrieve context from previous sessions, solving the problem of scattered knowledge across different tools. With its specialized 'Robot Mode' for machine-readable JSON output, token budget management, and multi-machine sync capabilities, CASS enables seamless knowledge transfer and context retrieval, making it easier for agents to learn from past interactions and solve complex debugging tasks.
主要功能
01Token budget management to truncate outputs and control LLM context usage
02Unified search across 11+ agents including Claude Code, Cursor, Aider, and Codex
03Robot Mode providing JSON/JSONL output optimized for AI agent consumption
04Multi-machine support via SSH/rsync to aggregate history from remote sources
050 GitHub stars
06Hybrid search modes combining lexical BM25 matching with vector similarity
使用场景
01Reviewing daily activity timelines to track development progress across multiple AI tools
02Searching past sessions across different agents to find previously implemented solutions
03Managing agent-to-agent handoffs by extracting summaries of previous work from different environments
