session-indexer

Session-Indexer is an innovative open-source tool designed for developers and AI practitioners who utilize Claude Code sessions. It offers per-project semantic search capabilities by indexing session transcripts stored as JSONL files into a local SQLite database. By leveraging advanced embedding techniques with Ollama's bge-m3 models and fallback to FTS5 BM25, it ensures efficient retrieval of relevant past interactions. The tool automatically injects pertinent context at the start of each session, enabling users to seamlessly pick up where they left off or review key decisions made earlier. Unlike shared memory stores, Session-Indexer maintains complete project privacy, making it ideal for sensitive or proprietary work. Its simplicity—no deployment required and easy installation—combined with robust features, makes it a compelling choice for developers aiming to enhance their workflow with intelligent session management.