Quickstart • MCP Server • AGENTS.md • CLI • Benchmarks
Semble is a code search library built for agents. It returns the exact code snippets they need instantly, using ~98% fewer tokens than grep+read. Indexing and searching a full codebase end-to-end takes under a second, with ~200x faster indexing and ~10x faster queries than a code-specialized transformer, at 99% of its retrieval quality (see benchmarks). Everything runs on CPU with no API keys, GPU, or external services. Run it as an MCP server or call it from the shell via AGENTS.md and any agent (Claude Code, Cursor, Codex, OpenCode, etc.) gets instant access to any repo.
Your agent queries Semble in natural language (e.g. "How is authentication handled?") and gets back only the relevant code snippets, without grepping or reading full files.
Semble has three complementary setup paths. The recommended setup is using all three (but you can pick and choose based on your needs):
- MCP server: an MCP server for your agent.
- AGENTS.md: an AGENTS.md snippet with instructions for calling Semble via the CLI.
- Sub-agent: a dedicated
semble-searchsub-agent for harnesses that support it.
Expose Semble as a native tool via MCP so your agent can call it directly. Add it to Claude Code (requires uv):
claude mcp add semble -s user -- uvx --from "semble[mcp]" sembleSee MCP Server below for other harnesses (Cursor, Codex, OpenCode, etc.).
Add Semble usage instructions to your agent's context so it knows when and how to call the CLI. Install the Semble CLI, then add the snippet below to your AGENTS.md or CLAUDE.md:
uv tool install semble # Install with uv (recommended)
pip install semble # Or with pipAGENTS.md / CLAUDE.md snippet
## Code Search
Use `semble search` to find code by describing what it does or naming a symbol/identifier, instead of grep:
```bash
semble search "authentication flow" ./my-project
semble search "save_pretrained" ./my-project
semble search "save model to disk" ./my-project --top-k 10
```
If you anticipate doing more than one search, use `semble index` to create an index.
```bash
semble index ./my-project -o my_index
```
You can then reuse this index later on:
```bash
semble search "save_pretrained" --index my_index
```
An index is not automatically updated, so if the code changes significantly, reindex. If you notice stale results while resolving searches to files, reindex.
Use `--content docs` to search documentation and prose, `--content config` for config files (yaml, toml, etc.), or `--content all` to search code, docs, and config:
```bash
semble search "deployment guide" ./my-project --content docs
semble search "database host port" ./my-project --content config
semble search "authentication" ./my-project --content all
```
Use `semble find-related` to discover code similar to a known location (pass `file_path` and `line` from a prior search result):
```bash
semble find-related src/auth.py 42 ./my-project
```
Like search, `find-related` also accepts an `--index` argument.
`path` defaults to the current directory when omitted; git URLs are accepted.
If `semble` is not on `$PATH`, use `uvx --from "semble[mcp]" semble` in its place.
### Workflow
1. Index the repo using `semble index -o cached_index`.
2. Start with `semble search` to find relevant chunks. Pass the index to achieve results faster.
3. Use `--content docs` for documentation, `--content config` for config files, or `--content all` for everything.
4. Inspect full files only when the returned chunk does not give enough context.
5. Optionally use `semble find-related` with a promising result's `file_path` and `line` to discover related implementations.
6. Use grep only when you need exhaustive literal matches or quick confirmation of an exact string.For harnesses that support sub-agents, install a dedicated semble-search sub-agent so search runs in its own context (requires the CLI):
semble init # Claude Code → .claude/agents/semble-search.mdSee Sub-agent setup below for other harnesses (Cursor, Codex, OpenCode, etc.).
Updating Semble
uv tool upgrade semble # with uv
uv cache clean semble # for MCP users (restart your MCP client after)
pip install --upgrade semble # with pip- Fast: indexes an average repo in ~250 ms and answers queries in ~1.5 ms, all on CPU.
- Accurate: NDCG@10 of 0.854 on our benchmarks, on par with code-specialized transformer models, at a fraction of the size and cost.
- Token-efficient: returns only the relevant chunks, using ~98% fewer tokens than grep+read.
- Zero setup: runs on CPU with no API keys, GPU, or external services required.
- MCP server: works with Claude Code, Cursor, Codex, OpenCode, VS Code, and any other MCP-compatible agent.
- Local and remote: pass a local path or a git URL.
Semble can run as an MCP server so agents can search any codebase directly. Repos are cloned and indexed on demand, and indexes are cached for the lifetime of the session. Local paths are watched for file changes and re-indexed automatically.
Requires uv to be installed.
Claude Code
claude mcp add semble -s user -- uvx --from "semble[mcp]" sembleCursor
Add to ~/.cursor/mcp.json (or .cursor/mcp.json in your project):
{
"mcpServers": {
"semble": {
"command": "uvx",
"args": ["--from", "semble[mcp]", "semble"]
}
}
}Codex
Add to ~/.codex/config.toml:
[mcp_servers.semble]
command = "uvx"
args = ["--from", "semble[mcp]", "semble"]OpenCode
Add to ~/.opencode/config.json:
{
"mcp": {
"semble": {
"type": "local",
"command": ["uvx", "--from", "semble[mcp]", "semble"]
}
}
}VS Code
Add to .vscode/mcp.json in your project (or your user profile's mcp.json):
{
"servers": {
"semble": {
"command": "uvx",
"args": ["--from", "semble[mcp]", "semble"]
}
}
}GitHub Copilot CLI
Add to ~/.copilot/mcp-config.json:
{
"mcpServers": {
"semble": {
"command": "uvx",
"args": ["--from", "semble[mcp]", "semble"]
}
}
}Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"semble": {
"command": "uvx",
"args": ["--from", "semble[mcp]", "semble"]
}
}
}Gemini CLI
Add to ~/.gemini/settings.json:
{
"mcpServers": {
"semble": {
"command": "uvx",
"args": ["--from", "semble[mcp]", "semble"]
}
}
}Kiro
Add to ~/.kiro/settings/mcp.json (or .kiro/settings/mcp.json in your project):
{
"mcpServers": {
"semble": {
"command": "uvx",
"args": ["--from", "semble[mcp]", "semble"]
}
}
}Zed
Add to ~/.config/zed/settings.json (or .zed/settings.json in your project):
{
"context_servers": {
"semble": {
"command": "uvx",
"args": ["--from", "semble[mcp]", "semble"]
}
}
}| Tool | Description |
|---|---|
search |
Search a codebase with a natural-language or code query. Pass repo as a local directory path or an https:// git URL. |
find_related |
Given a file path and line number, return chunks semantically similar to the code at that location. |
By default the MCP server indexes only code files. To also index documentation, config, or everything, append --content docs, --content config, or --content all to the server command, or a combination, e.g. --content code docs. For example, in Claude Code: claude mcp add semble -s user -- uvx --from "semble[mcp]" semble --content all.
Claude Code, Gemini CLI, Cursor, OpenCode, GitHub Copilot CLI, and Kiro all support a dedicated semble search sub-agent. Run semble init once in your project root:
semble init # Claude Code → .claude/agents/semble-search.md
semble init --agent gemini # Gemini CLI → .gemini/agents/semble-search.md
semble init --agent cursor # Cursor → .cursor/agents/semble-search.md
semble init --agent opencode # OpenCode → .opencode/agents/semble-search.md
semble init --agent copilot # Copilot CLI → .github/agents/semble-search.md
semble init --agent kiro # Kiro → .kiro/agents/semble-search.mdIf semble is not on $PATH, prefix the command with uvx --from "semble[mcp]".
Semble also ships as a standalone CLI. This is useful in scripts or anywhere you want search results without an MCP session.
# Search a local repo
semble search "authentication flow" ./my-project
# Index first for faster repeated searches (--index works with any command below)
semble index ./my-project -o my-index
semble search "authentication flow" --index my-index
# Search a remote repo (cloned on demand)
semble search "save model to disk" https://github.com/MinishLab/model2vec
# Limit results
semble search "save model to disk" ./my-project --top-k 10
# Search docs/config/everything instead of just code
semble search "deployment guide" ./my-project --content docs # or: config, all
# Find code similar to a known location
semble find-related src/auth.py 42 ./my-project--content accepts code (default), docs, config, or all. path defaults to the current directory when omitted; git URLs are accepted. If semble is not on $PATH, use uvx --from "semble[mcp]" semble in its place.
Savings
semble savings shows how many tokens semble has saved across all your searches:
semble savings # summary by period
semble savings --verbose # also show breakdown by call type Semble Token Savings
════════════════════════════════════════════════════════════════
Period Calls Savings
────────────────────────────────────────────────────────────────
Today 42 [███████████████░] ~58.4k tokens (95%)
Last 7 days 287 [██████████████░░] ~312.4k tokens (90%)
All time 1.4k [██████████████░░] ~1.2M tokens (89%)
Savings are calculated as follows: for each call, semble records the total character count of the unique files containing returned chunks and the character count of the snippets returned. Estimated tokens saved is (file chars − snippet chars) / 4 (4 chars per token). This is a conservative estimate: the baseline is reading matched files in full, which is how coding agents often explore unfamiliar code.
Stats are stored in ~/.semble/savings.jsonl.
Library usage
Semble can also be used as a Python library for programmatic access, useful when building custom tooling or integrating search directly into your own code.
from semble import ContentType, SembleIndex
# Index a local directory (code only, the default)
index = SembleIndex.from_path("./my-project")
# Index docs and prose (markdown, rst, etc.)
index = SembleIndex.from_path("./my-project", content=ContentType.DOCS)
# Index everything (code, docs, and config)
index = SembleIndex.from_path("./my-project", content=[ContentType.CODE, ContentType.DOCS, ContentType.CONFIG])
# Index code and docs together
index = SembleIndex.from_path("./my-project", content=[ContentType.CODE, ContentType.DOCS])
# Index a remote git repository
index = SembleIndex.from_git("https://github.com/MinishLab/model2vec")
# Search the index with a natural-language or code query
results = index.search("save model to disk", top_k=3)
# Find code similar to a specific result
related = index.find_related(results[0], top_k=3)
# Each result exposes the matched chunk
result = results[0]
result.chunk.file_path # "model2vec/model.py"
result.chunk.start_line # 127
result.chunk.end_line # 150
result.chunk.content # "def save_pretrained(self, path: PathLike, ..."We benchmark quality and speed across ~1,250 queries over 63 repositories in 19 languages (left), and token efficiency against grep+read at equivalent recall levels (right).
 |
 |
The quality benchmark (left) scores retrieval quality (NDCG@10) against total latency; semble achieves 99% of the quality of the 137M-parameter CodeRankEmbed Hybrid while indexing 218x faster. The token efficiency benchmark (right) measures how many tokens each method needs to reach a given recall level; semble uses 98% fewer tokens on average and hits 94% recall at only 2k tokens, while grep+read needs a full 100k context window to reach 85%. See benchmarks for per-language results, ablations, and full methodology.
Semble splits each file into code-aware chunks using tree-sitter, then scores every query against the chunks with two complementary retrievers: static Model2Vec embeddings using the code-specialized potion-code-16M model for semantic similarity, and BM25 for lexical matches on identifiers and API names. The two score lists are fused with Reciprocal Rank Fusion (RRF).
After fusing, results are reranked with a set of code-aware signals:
Ranking signals
- Adaptive weighting. Symbol-like queries (
Foo::bar,_private,getUserById) get more lexical weight, while natural-language queries stay balanced between semantic and lexical retrievers. - Definition boosts. A chunk that defines the queried symbol (a
class,def,func, etc.) is ranked above chunks that merely reference it. - Identifier stems. Query tokens are stemmed and matched against identifier stems in a chunk, giving an additional weight to chunks that contain them. For example, querying
parse configboosts chunks containingparseConfig,ConfigParser, orconfig_parser. - File coherence. When multiple chunks from the same file match the query, the file is boosted so the top result reflects broad file-level relevance rather than a single out-of-context chunk.
- Noise penalties. Test files,
compat//legacy/shims, example code, and.d.tsdeclaration stubs are down-ranked so canonical implementations surface first.
Because the embedding model is static with no transformer forward pass at query time, all of this runs in milliseconds on CPU.
MIT
If you use Semble in your research, please cite the following:
@software{minishlab2026semble,
author = {{van Dongen}, Thomas and Stephan Tulkens},
title = {Semble: Fast and Accurate Code Search for Agents},
year = {2026},
publisher = {Zenodo},
doi = {10.5281/zenodo.19785932},
url = {https://github.com/MinishLab/semble},
license = {MIT}
}