HN Showcase
Back to browse
GitHub Repository

Local-first hybrid semantic code search tool. Indexes codebases into PostgreSQL with pgvector embeddings via Ollama, combines vector similarity + keyword search with RRF fusion. Supports 30+ languages. Features CLI, MCP server, WEB dashboard and interactive REPL.

32 starsPython

CocoSearch – semantic code search with syntax-aware chunking

by VioletCranberry·Feb 28, 2026·2 points·1 comment
Visit ProjectView on HN

AI Analysis

●●SolidSolve My ProblemDark Horse

Hybrid vector+keyword search beats single-mode retrieval, but Cursor and Cody already own this.

Strengths
  • •Syntax-aware chunking via tree-sitter preserves semantic boundaries across 30+ languages
  • •Runs fully local with Ollama—no vendor lock-in or API costs for embeddings
  • •RRF fusion + keyword fallback handles both semantic intent and exact term matching
Weaknesses
  • •Crowded space: Cursor, Continue, Sourcegraph Cody already solve code search at scale
  • •PostgreSQL + Ollama + pgvector setup barrier higher than web SaaS alternatives
Target Audience

Backend developers, engineering teams managing large codebases

Similar To

Cursor · Sourcegraph Cody · Continue.dev

Similar Projects

Developer Tools●●●Banger

Srclight – Deep code indexing MCP server (FTS5 and Tree-sitter)

Tree-sitter + FTS5 + MCP = tokens saved for AI agents to actually code, not search.

WizardryBig BrainZero to One
srclight
103mo ago