服务器
C
ContextBridge
@tijuthomas5
Local-first code retrieval for AI agents — cuts codebase context from tens of thousands of tokens to a few hundred, with zero hallucinated file paths.
ContextBridge sits between your codebase and your AI coding agent. Instead of pasting raw source files into context (burning thousands of tokens on code that isn't relevant) or letting the AI guess which files matter, CB indexes your codebase's real structure and returns a compact, ranked result: the owner file, related files, key symbols, and dependency chains — grounded in your actual code, not guessed.
- ~96% token reduction on typical queries (measured: 118 KB raw context → 4–8 KB grounded result)
- Zero hallucinated paths — every file/method returned is verified to exist, not inferred from naming patterns
- Local-first — your code never leaves your machine; retrieval and optional analysis both run locally
- Works with any AI client — Claude Code, Cursor, Codex, Antigravity, Gemini CLI, or anything that speaks MCP, over SSE or stdio
- Generic engine — no project-specific tuning baked in; a swappable profile plugin handles domain-specific ranking, so it works on any codebase
- Three retrieval modes — hybrid (keyword + vector), semantic (vector-only), keyword (fast exact lookups)
