TL;DR
Semble is a code search tool designed for agents that uses approximately 98% fewer tokens than traditional grep methods. It offers rapid, accurate code retrieval on CPU without external dependencies. This development could significantly improve code search efficiency for AI agents and developers.
Semble, a new code search library tailored for AI agents, claims to reduce token usage by approximately 98% compared to traditional grep-based searches, while delivering faster, accurate results on CPU without external dependencies.
Semble is designed to enable agents such as Claude Code, Codex, and OpenCode to perform instant, precise code searches within large codebases. It indexes repositories in around 250 milliseconds and answers queries in roughly 1.5 milliseconds, all on CPU, with no need for API keys or external services. Benchmarks indicate its retrieval quality is comparable to specialized transformer models, but with significantly lower resource requirements.
The tool can be integrated as an MCP server or used via command-line, supporting local paths and git URLs. It provides features like natural language search queries (e.g., ‘How is authentication handled?’) and related code discovery, making it a practical enhancement for code exploration in AI workflows.
Why It Matters
This development is important because it offers a highly efficient, cost-effective way for AI agents to access and understand codebases, potentially accelerating development workflows and reducing computational costs. Its token efficiency and speed could make code search more accessible and scalable for large projects and multiple users.
FOXWELL NT301 OBD2 Scanner Live Data Professional Mechanic OBDII Diagnostic Code Reader Tool for Check Engine Light
【Vehicle CEL Doctor】The NT301 obd2 scanner enables you to read DTCs, access to e-missions readiness status, turn off…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background
Traditional code search methods like grep read entire files, consuming many tokens and time, especially in large repositories. Existing transformer-based models provide accurate search results but are resource-intensive. Semble aims to bridge this gap by offering comparable accuracy with minimal token usage and faster performance, addressing a key bottleneck in AI-assisted coding workflows.
“Semble returns only the relevant chunks, using approximately 98% fewer tokens than grep+read, while maintaining high retrieval quality.”
— Semble developers
“Indexing and searching a full codebase takes under a second, which is a significant improvement for developers and AI agents alike.”
— Hacker News user
AI Data Analysis for Non-Coders: Use Claude Code to Clean Spreadsheets, Build Dashboards, and Automate Reports — The Vibe Coder's Handbook (Book 4)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What Remains Unclear
It is not yet clear how Semble performs across a wide variety of codebases or in different programming languages beyond initial benchmarks. Long-term stability and integration support are still to be tested in diverse environments.
Code by Note, Bk 1: Find the Patterns by Reading the Notes, Coloring Book (Color by Note, Bk 1)
Used Book in Good Condition
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What’s Next
Next steps include broader adoption by AI developers, further benchmarking across various repositories, and potential integration into more agent workflows. Updates may include feature enhancements and expanded language support.
Generative AI for Software Development: Building Software Faster and More Effectively
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
How does Semble compare in accuracy to existing code search tools?
Benchmarks show Semble achieves approximately 99% of the retrieval quality of specialized transformer models, with significantly lower token usage and faster response times.
Can Semble be used with any codebase or programming language?
Semble supports local repositories and git URLs, but detailed language support is not specified. Its performance in languages other than those tested remains to be confirmed.
Is Semble easy to set up and integrate into existing workflows?
Yes, it runs on CPU with no external dependencies. It can be added as an MCP server or called via CLI, with straightforward setup instructions for popular agents like Claude Code, Codex, and OpenCode.
What are the limitations or potential drawbacks of Semble?
Its performance across diverse codebases and in different languages is still being evaluated. Long-term stability and compatibility with various agent architectures are also to be determined.
