Show HN: Zerikai Memory: local code-memory MCP server, now with Jev

Category: ai-ml

Tags: mcp, code-indexing, local-first, vector-database, ai-assistant, context-management

Score: 7.3/10 (Innovation: 8, Technical: 7, Documentation: 7, Utility: 7)

Zerikai Memory is a local-first MCP server that provides persistent, workspace-isolated memory for AI coding assistants by indexing codebases with tree-sitter and ChromaDB, then retrieving relevant context to reduce token costs. It now integrates Jev, an AI judgment layer that filters and scores retrieved passages for relevance and prompt injection, enhancing retrieval quality. The project addresses a real pain point of context loss in AI-assisted development and offers a cost-aware solution.

Target audience: AI developers, software engineers using AI coding assistants, DevOps and platform engineers

Repository: https://github.com/KikeVen/zerikai_memory · Python · MIT · 36 stars

View on Hacker News