Skip to main content
Glama
chunkhound

ChunkHound

by chunkhound
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

Requirements

  • Python 3.10+

  • uv — install via curl -LsSf https://astral.sh/uv/install.sh | sh

  • API keys (optional — regex search works without any):

Related MCP server: OpenCodeHub MCP Server

AI writes code blind

Agents can generate code, but they still miss the context that makes software safe to change: how behavior flows across files, what changed across a branch or release, and which external constraints matter.

Reviewers, support, and product teams hit the same wall when large PRs, merge conflicts, bugs, and release notes need implementation-backed explanation instead of guesses.

ChunkHound turns current code, git history, and technical web research into cited context before anyone edits, reviews, debugs, or explains software.

Deep understanding for four context-heavy jobs

ChunkHound applies codebase understanding to the workflows where missing context hurts most.

Research before editing

Give coding agents grounded architecture context, relevant files, recent changes, and external constraints before they write code.

Understand large PRs and releases

Turn branch diffs, commit ranges, tags, and specific commits into cited engineering briefs for review, release notes, and changelog drafts.

Trace bugs and incidents

Turn symptoms, stack traces, and customer reports into likely code paths, recent changes, and external constraints.

Reconcile code with external docs

Pinpoint the technical docs, APIs, issues, and articles your implementation depends on, then connect that external evidence to local code research.

What you can ask

Ground an agent before edits

chunkhound research "How does authentication work?"
chunkhound search "JWT refresh token validation"
chunkhound research "What changed in auth recently?" --last-n 20

Understand a large PR or release

chunkhound research "Summarize the behavior changes on this branch for reviewers" --commit-range main..HEAD
chunkhound research "Draft changelog bullets for billing since v2.4" --commit-range v2.4..HEAD
chunkhound search "database migration" --commit-hash abc1234

Get context before resolving conflicts

chunkhound research "Why did auth session handling change on each side?" --commit-range main..feature/auth
chunkhound search "session refresh conflict" --last-n 50

Trace a bug with external constraints

chunkhound research "why would webhook retries fail?"
chunkhound research "what changed in webhook handling this week?" --last-n 30
chunkhound websearch "Stripe webhook retry schedule"

Explain product behavior

chunkhound research "What happens when a user cancels a subscription?"
chunkhound research "What changed in billing since v2.4?" --commit-range v2.4..HEAD

What powers deep understanding

  • Semantic code search — find relevant code by meaning, not only exact text

  • Cited code research — explain behavior across files with source citations

  • Git history research — ask by last N commits, commit hash, tag, branch, or range to understand large PRs and releases

  • Pinpoint web research — bring cited external docs, APIs, issues, and articles into the same workflow as local code research

  • Autodoc — generate shareable docs from code-backed research

  • Local-first indexing — keep code search and indexing under your control

  • Python, JavaScript, TypeScript, Java, Go, Rust, C/C++, and more via Tree-sitter

Install

uv tool install chunkhound

Try it

chunkhound index .
chunkhound research "How does authentication work?"

Index once, ask a real architecture question, and get a grounded answer with citations. Regex search works without providers. Semantic search requires an embedding provider. Deep research requires an LLM provider and an embedding provider with reranking support; web research uses the same provider stack. Choose local providers for zero-code-egress setups.

For a full configurable setup, create .chunkhound.json in your project root:

{
  "embedding": { "provider": "voyageai", "api_key": "your-key" },
  "llm": { "provider": "claude-code-cli" }
}

For editor integration, all provider options, and advanced configuration:

→ chunkhound.ai/docs/getting-started

Search git history

In addition to searching your indexed codebase, ChunkHound can search code changes across git history — useful for understanding what changed in a PR, a release, or since a specific commit.

# Last N commits
chunkhound search "authentication changes" --last-n 20

# Changes introduced by a specific commit
chunkhound search "database migration" --commit-hash abc1234

# Custom git range
chunkhound search "API changes" --commit-range v2.0..HEAD

# Deep research over recent changes
chunkhound research "what changed in the auth module?" --last-n 50

--vector-source controls scope: diff (default, changed code only), both (merges diff + DB), db (ignore diff).

Good fit

ChunkHound is especially useful for:

  • large repos and monorepos

  • multi-language codebases

  • legacy systems

  • local-only or security-sensitive environments

  • engineering teams that want agents, support, and product questions grounded in the same code index

Community

ChunkHound is MIT licensed, open source, and community built.

License

MIT

Available Tools

2 tools
daemon_statusA

Report daemon startup, scan, and realtime indexing health.

USE FOR:

  • Checking whether initial indexing has completed

  • Inspecting backend-neutral realtime health and resync state

  • Debugging degraded daemon behavior without opening log files

OUTPUT: {status, query_ready, scan_progress}

ParametersJSON Schema
NameRequiredDescriptionDefault

No parameters

TDQS

A4.6/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

No annotations provided, but the description discloses the tool's behavior: it reports health status without side effects. It outlines the output fields, though it could explicitly state it is read-only.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is concise and structured with clear sections (USE FOR, OUTPUT), no wasted words, and efficient communication of purpose and usage.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given no parameters and no output schema, the description adequately covers what the tool does and its output structure. It could mention potential error conditions, but it is fairly complete for the tool's simplicity.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

There are zero parameters and schema coverage is 100%, so the description does not need to add parameter information. The baseline for no parameters is 4.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states it reports daemon startup, scan, and realtime indexing health, with specific use cases and output fields. It distinguishes from sibling 'search' which is unrelated.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The 'USE FOR' section explicitly lists three specific scenarios (checking indexing completion, inspecting realtime health, debugging without logs), providing clear guidance on when to use this tool.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 2 tool updatesv0.1.0
    • First observeddaemon_status
    • First observedsearch
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A4.1/5.0

Scored across 2 tools

Disambiguation5/5

Each tool has a clearly distinct purpose: daemon_status for health monitoring and search for code retrieval. No overlap in functionality.

Naming Consistency5/5

Both tool names use consistent snake_case format: daemon_status and search. Naming pattern is uniform.

Tool Count3/5

With only 2 tools, the server feels minimal for a code indexing service. While functional, it lacks typical tools like index management or configuration, placing it below the ideal range.

Completeness2/5

The tool surface is severely incomplete for code indexing: missing index control, rescan triggers, configuration, or result filtering. Only basic search and status checks are covered.

Maintenance

ActivityActive
ResponsivenessResponsive

Resources

Related MCP Connectors

Related MCP Servers

  • agentmako

    drhalto
    A
    license
    C
    quality
    A
    maintenance
    Local-first codebase intelligence engine providing AI coding agents with a typed MCP toolset for understanding and navigating code repositories.
    100
    51
    Apache 2.0

  • OpenCodeHub MCP Server

    laithalsaadoon
    A
    license
    Not graded
    quality
    A
    maintenance
    Provides code intelligence for AI coding agents by indexing repositories into a hybrid knowledge graph, enabling agents to query dependencies, impact, and context through 28 MCP tools.
    3
    Apache 2.0

  • Paparats MCP

    IBazylchuk
    A
    license
    Not graded
    quality
    B
    maintenance
    Provides AI coding assistants with deep, semantic understanding of local codebases via AST-aware chunking, cross-repo symbol graphs, and architectural memory, enabling context-aware code search and dependency tracing.
    10
    MIT

  • Atlas

    Brair-Mpagi
    A
    license
    Not graded
    quality
    B
    maintenance
    A local-first codebase intelligence layer for AI coding agents, providing a persistent, queryable model of a repository via an MCP server and CLI to enable structure queries instead of reading many files.
    Apache 2.0

Appeared in Searches