Chroma Package Search

@chroma-core

About Chroma Package Search

Give coding agents the ability to understand dependencies.

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "package-search-mcp": {
      "type": "http",
      "url": "https://mcp.trychroma.com/package-search/v1",
      "headers": {
        "x-chroma-token": "<CHROMA_API_TOKEN>"
      }
    }
  }
}

Tools

## Description Reads exact lines from a source file of a public package. Useful for fetching specific code regions by line range. Required Args: "registry", "package_name", "filename_sha256", "start_line", "end_line" Optional Args: "version" Best for: Inspecting exact code snippets when you already know the file and line numbers. Max 200 lines.

## Description Searches package source code using semantic understanding AND optionally regex patterns. This allows for hybrid search, allowing for prefiltering with regex, and semantic ranking. Required Args: "registry", "package_name", "semantic_queries" Optional Args: "version", "filename_sha256", "pattern", "language" Best for: Understanding how packages implement specific features, finding usage patterns, or exploring code structure.

## Description Executes a grep over the source code of a public package. This tool is useful for deterministically finding code in a package using regex. Use this tool before implementing solutions that use external packages. The regex pattern should be restrictive enough to only match code you're looking for, to limit overfetching. Required Args: "registry", "package_name", "pattern" Optional Args: "version", "language", "filename_sha256", "a", "b", "c", "head_limit", "output_mode" Best for: Deterministic code search, finding specific code patterns, or exploring code structure.

Overview

What is Chroma Package Search?

Chroma Package Search is an MCP server that adds ground truth context about code packages to AI agents. It exposes regex and semantic query capabilities over popular open source code packages to improve coding task performance and reduce hallucination.

How to use Chroma Package Search?

Obtain an API key from Chroma, then invoke the server’s three tools: package_search_grep (regex grep), package_search_hybrid (semantic + regex), and package_search_read_file (read specific line ranges). The API key can be acquired at https://trychroma.com/package-search.

Key features of Chroma Package Search

Deterministic regex search over package source code
Hybrid search combining semantic understanding and regex patterns
Read exact lines from source files by line range
Improves AI agent coding task performance
Reduces hallucination with ground truth context

Use cases of Chroma Package Search

AI assistants verifying package code behavior
Code generation with accurate dependency context
Debugging with access to package source snippets

FAQ from Chroma Package Search

What tools does Chroma Package Search provide?

It provides three tools: package_search_grep for regex search, package_search_hybrid for semantic+regex search, and package_search_read_file for reading specific line ranges.

How do I get an API key?

Create a Chroma account and get an API key at https://trychroma.com/package-search. You receive $5 worth of free credits, roughly 1,000 semantic queries.

What is the benefit of using this server?

By exposing the source code of a project’s dependencies to a model, it improves performance on coding tasks and reduces the potential for hallucination.

Frequently asked questions

What tools does Chroma Package Search provide?

It provides three tools: `package_search_grep` for regex search, `package_search_hybrid` for semantic+regex search, and `package_search_read_file` for reading specific line ranges.

How do I get an API key?

Create a Chroma account and get an API key at https://trychroma.com/package-search. You receive $5 worth of free credits, roughly 1,000 semantic queries.

What is the benefit of using this server?

By exposing the source code of a project’s dependencies to a model, it improves performance on coding tasks and reduces the potential for hallucination.

Comments

Basic information

More Search MCP servers

mcp-omnisearch

spences10

🔍 A Model Context Protocol (MCP) server providing unified access to multiple search engines (Tavily, Brave, Kagi, Exa), AI tools (Kagi FastGPT, Exa, Linkup), and content extraction services (Firecrawl, Tavily, Kagi). Includes GitHub search. All through a single interface.

332

Web Scout MCP Server

pinkpixel-dev

A powerful MCP server extension providing web search and content extraction capabilities. Integrates DuckDuckGo search functionality and URL content extraction into your MCP environment, enabling AI assistants to search the web and extract webpage content programmatically.

129