What is Ariadne and its core purpose?

Ariadne is a semantic code index and gatekeeper for AI coding agents and developers. It uses static analysis to build a compiler-grade symbol and dependency graph of your codebase, enabling powerful queries and ensuring repository consistency.

How does Ariadne enhance AI coding agents like Claude Code?

Ariadne integrates seamlessly with AI agents by injecting pre-edit context (impact analysis), performing post-edit updates, logging post-test runs, and acting as a stop-time verification gate to ensure code quality and prevent issues from being committed.

Which programming languages does Ariadne support for indexing?

Ariadne offers comprehensive indexing support for multiple languages including TypeScript, Python, Go, Rust, Java/Kotlin, Scala, C#, and Ruby, along with non-code artifacts like `.env` and `package.json`.

What kind of analysis and queries can I perform with Ariadne?

You can perform advanced queries such as searching symbols, jumping to definitions, finding all references, tracing the impact of changes, exploring call graphs, and visualizing file or module dependency graphs via CLI or an MCP server.

Ariadne: Semantic Code Indexing & AI Agent Gatekeeper

Ariadne is a sophisticated semantic code index and gatekeeper designed to empower AI coding agents without relying on large language models. It achieves this by performing compiler-grade static analysis to build a detailed symbol graph, tracking definitions, references, and file-level dependencies through import extraction across various programming languages. This robust indexing allows for advanced queries, enabling developers and AI agents to search symbols, trace change impacts, explore call graphs, verify code consistency, and even index non-code artifacts, ensuring a comprehensive understanding and control over the codebase. It integrates via CLI, MCP server tools, and specific AI agent hooks.

主な機能

01Indexes codebases in multiple languages (TypeScript, Python, Go, Rust, Java/Kotlin, Scala, C#, Ruby) into a SQLite graph of symbols, references, and dependencies.

02Offers powerful CLI and MCP server queries for searching symbols, jumping to definitions, finding references, performing impact analysis, and exploring call and module graphs.

03Verifies repository consistency through gatekeeper checks for index freshness, missing test runs, type errors, and architecture boundary violations.

04Indexes non-code artifacts such as `.env` variables, `package.json` scripts, SQL migrations, and OpenAPI schemas, linking them to source code references.

05Integrates seamlessly with AI coding agents like Claude Code via pre-edit context injection, post-edit updates, post-test logging, and a stop-time verification gate.

060 GitHub stars

ユースケース

01Enhancing AI coding agent capabilities by providing deep, compiler-grade code context and automated verification checks during development cycles.

02Empowering developers to perform advanced code exploration, impact analysis, and dependency tracing across large codebases or monorepos.

03Maintaining codebase health and enforcing architectural standards through automated consistency checks before code is finalized.