← The Wire

Tools2026-06-08 · source-backed

zilliztech/claude-context cuts retrieval tokens ~40% with hybrid code search.

Confidence
source-backed
Sources
1
Redaction
redacted

Story

claude-context (11.8K stars) is an MCP server that indexes your whole codebase into Milvus and exposes hybrid BM25 + dense-vector search to Claude Code and any MCP agent. Ask "find functions that handle user authentication" instead of grepping. Zilliz reports ~40% token reduction at equal retrieval quality, with support for OpenAI, VoyageAI, Ollama, and Gemini embeddings. On a large repo this is the difference between the agent reading three relevant functions and dumping whole files into context.

Related stories

Each link below shares sources, entities, or timing with this story.

  1. Shared entities / Same source domain / Shared topic / Earlier coverage

    10 Skills of the Day

    Both cover BM25, Claude Code, MCP; reported by the same outlet (github.com); overlapping topics (agent, between, claude, code, context).

  2. Cherry Studio Crosses 45.3K Stars: Open-Source AI Desktop Client With 300+ Assistants.

    Both cover Gemini, MCP, Ollama, OpenAI; reported by the same outlet (github.com); overlapping topics (agent, between).

  3. Shared entities / Same source domain / Shared topic / Earlier coverage / Tension

    The Claude Code Field Guide Hit 53,400 Stars. Here's Why That Matters.

    Both cover Claude Code, Gemini, MCP; reported by the same outlet (github.com); overlapping topics (agent, claude, code).

  4. Shared entities / Same source domain / Shared topic / What happened next

    AI coding tools converge on portable skills

    Both cover Claude Code, Gemini, MCP; reported by the same outlet (github.com); overlapping topics (agent, claude, code).

  5. Shared entities / Same source domain / Shared topic / Earlier coverage

    Skills You Can Learn Today

    Both cover Claude Code, Gemini, MCP; reported by the same outlet (github.com); overlapping topics (agent, claude, code).

  6. Shared entities / Same source domain / Shared topic / What happened next

    A 54k-star skill that makes Claude talk like a caveman, and the math that deflates it

    Both cover Claude Code, Gemini; reported by the same outlet (github.com); overlapping topics (agent, claude, code, context, token).

  7. Shared entities / Same source domain / Shared topic / Earlier coverage

    rtk hits 59.7K stars: a zero-dependency Rust proxy that cuts agent token use 60 to 90%

    Both cover Claude Code, MCP; reported by the same outlet (github.com); overlapping topics (agent, claude, code, token, whole).

  8. lean-ctx Ships "Context OS" for AI Development. 60-99% Token Reduction via MCP Server.

    Both cover Claude Code, MCP; reported by the same outlet (github.com); overlapping topics (between, claude, code, context, token).

Source trail

Entities

Provenance

AI generated
no
Story unit
2026-06-08-zilliztech-claude-context-cuts-retrieval-tokens-40-with-hybrid-code-search
Labels
source-backed, canonical briefing excerpt