# MinishLab/semble

Fast and Accurate Code Search for Agents. Uses 99% fewer tokens than grep+read

Repository: https://github.com/MinishLab/semble
Canonical: https://ross.abutalabs.com/products/semble
Homepage: https://minish.ai/packages/semble/introduction/
Language: Python
License: MIT
License Family: permissive
Topics: agents, code-search, embeddings, mcp, mcp-server, model-context-protocol, retrieval
Last push: 2026-08-26T04:48:21+00:00

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 97, longevity 10
- inputs: {"age_days": 149, "days_push": 7, "days_rel": 21, "gap_med": 3, "n_releases_24m": 26}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5958, forks 257 (observed 2026-08-28T04:09:33.672508+00:00)

## What it is
Semble is a fast, CPU-only code search library built for coding agents, providing semantic and hybrid search over codebases with ~99% fewer tokens than grep+read. It integrates as an MCP server, CLI tool via AGENTS.md instructions, or a dedicated sub-agent for agents like Claude Code, Cursor, and Codex.

## Use cases
- let coding agents search a codebase semantically instead of grepping
- reduce token usage when agents explore large repositories
- index and query a repo in under a second on CPU
- expose code search as an MCP tool for Claude Code or Cursor
- find how authentication or a feature is implemented via natural language queries
- search remote git repositories without cloning manually

## When to choose
- you use AI coding agents and want fast, accurate, token-efficient code retrieval
- you need code search that runs fully on CPU with no API keys or external services
- you want an MCP server or sub-agent integration for code search

## When to avoid
- you need GPU-accelerated or transformer-grade retrieval beyond its benchmark quality
- you need a hosted or cloud-based code search service
- your project is not Python 3.10+ compatible

## Facets
- artifact type: library
- maturity: active
- function: search-engine, rag, mcp, cli, nlp
- domain: developer-tools, artificial-intelligence, large-language-models
- platform: python, cli, cross-platform
- tags: code-search, semantic-search, hybrid-search, embeddings, mcp-server, coding-agents, token-efficiency, cpu-only, ai-agents, search, docker

## Member repositories
- MinishLab/semble (main) score 80

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:33.672508+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:50:25.263872+00:00, confidence not recorded.
  - readme: https://github.com/MinishLab/semble (fetched 2026-08-28T04:09:33.672508+00:00, sha f7a7019a6de7)
  - homepage: https://minish.ai/packages/semble/introduction/ (fetched 2026-08-29T08:46:16.572378+00:00, sha 459ff1bacb11)
  - site_page: https://minish.ai/about (fetched 2026-08-29T08:46:16.575242+00:00, sha 6332b2896728)
  - site_page: https://minish.ai/packages/model2vec/installation (fetched 2026-08-29T08:46:16.577071+00:00, sha 21a9aeffe4ea)
  - site_page: https://minish.ai/packages/semble/installation (fetched 2026-08-29T08:46:16.580538+00:00, sha 373938e60d5f)
  - site_page: https://minish.ai/packages/semhash/installation (fetched 2026-08-29T08:46:16.582239+00:00, sha ad1ab7983910)
  - site_page: https://minish.ai/packages/vicinity/installation (fetched 2026-08-29T08:46:16.583915+00:00, sha 4ed119eab51b)
  - registry_pypi: https://pypi.org/pypi/semble/json (fetched 2026-08-29T08:46:16.585523+00:00, sha 4cabd801b269)
  - site_page: https://minish.ai/packages/model2vec/integrations (fetched 2026-08-29T08:46:16.578782+00:00, sha d24212c8593e)
- Data as of 2026-08-30T08:39:29.467469+00:00.