semble

Overview

Semble is a fast and accurate code search library designed specifically for AI agents. It returns the exact code snippets needed instantly, using significantly fewer tokens than traditional grep+read approaches. The solution indexes and searches a full codebase in under a second, with ~200x faster indexing and ~10x faster queries than code-specialized transformer models, while maintaining 99% of retrieval quality. Everything runs on CPU with no API keys, GPU, or external services required.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• search

  Search a codebase with a natural-language or code query.

  Try:How is authentication handled?Try:Find code related to saving models to diskTry:Search for 'save_pretrained' function
• find_related

  Given a file path and line number, return chunks semantically similar to the code at that location.

  Try:Find code similar to the authentication flow in src/auth.py at line 42Try:Show me code related to the implementation at utils.py 156

Comparable tools

Installation

pip install semble[mcp]
# or
uv tool install semble

{
  "mcpServers": {
    "semble": {
      "command": "uvx",
      "args": ["--from", "semble[mcp]", "semblable"]
    }
  }
}

[mcp_servers.semble]
command = "uvx"
args = ["--from", "semblable[mcp]", "semblable"]

FAQ

How does Semble compare to traditional grep+read?

Semble uses approximately 98% fewer tokens than grep+read by returning only the relevant code chunks rather than entire files.

What agents support Semble as an MCP server?

Semble works with Claude Code, Cursor, Codex, OpenCode, and any other MCP-compatible agent.

On Hacker News

Recent discussion from the developer community.

• Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep▲ 420💬 137
  Story by Bibabomas · 2026-05-17
• Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep▲ 8
  Story by stephantul · 2026-05-03
• Show HN: Semble – Fast code search for agents with near-transformer accuracy▲ 7
  Story by stephantul · 2026-04-26

Compare semble with