Nicheloom

Market intelligence for builders — see what's gaining traction before it's crowded.

Semble

Fast code search for agents with near-transformer accuracy

Details

External ID
47910885
Source
HN
Company
—
Product
Semble
Website domain
github.com
Launched
April 26, 2026
Cohort
—
Upvotes
7
Upvotes percentile
0.38817480719794345
Tags
—
Fetched at
Sept. 7, 2026, 9:26 p.m.
Updated at
Sept. 7, 2026, 9:26 p.m.

Description

Hey HN! We've just open-sourced Semble, a fast and accurate code search library built for agents. We're also releasing potion-code-16M, a small code-specialized static embedding model that powers it.Most embedding-based code search methods are either too slow to index on demand or need GPU infrastructure, while grep-style retrieval methods often cannot find the relevant content. Semble combines the speed and quality benefits of both, so agents waste less time and fewer tokens exploring.Main features:- Fast: indexes a full codebase in ~250 ms and answers queries in ~1.5 ms, all on CPU (roughly ~200x faster indexing and ~10x faster queries than a code-specialized transformer).- Accurate: on par with code-specialized transformer models at a fraction of the size (see our benchmarks for more info).- MCP server: drop-in tool for Claude Code, Cursor, Codex, OpenCode, and any other MCP-compatible CLI/agent. Repos are cloned and indexed on demand.- Zero setup: runs on CPU with no API keys, GPU, or external services.Install as an MCP server for Claude Code:claude mcp add semble -s user -- uvx --from "semble[mcp]" sembleOr check our README for install instructions for Codex, OpenCode, Cursor, and other agents.Semble: https://github.com/MinishLab/sembleBenchmarks: https://github.com/MinishLab/semble/tree/main/benchmarksHow it works: https://github.com/MinishLab/semble#how-it-worksModel: https://huggingface.co/minishlab/potion-code-16M

Enrichment

Theme
Claude integrations and coding agents
Vertical
—
Function
Search & retrieval
Audience
Developer
AI stance
AI feature
Project type
Commercial product
Normalized one-liner
fast code search for ai agents
Manually corrected
False

Could you build this?

Partial The client library and local search index can be vibe-coded, but training a specialized 16M parameter static embedding model from scratch requires custom ML training data curation and modeling.

What it would actually take: A full version requires training a specialized static embedding model (like potion-code-16M) on billions of AST/code tokens using contrastive learning and distillation techniques. The serving stack uses a high-performance vector index (written in Rust or C++) optimized for CPU cache efficiency and sub-millisecond search across local codebases.

Discussion

No comments on this launch.

Competitors

Other products that read as similar to this one — 214 launches clear the similarity bar, closest 8 shown.

Attention rank: #127 of 215 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).

Launched 170 days after the earliest competitor.

Other launches for this product