Sourcegraph Cody

Sourcegraph Cody is an AI coding assistant that differentiates itself through deep codebase understanding powered by Sourcegraph's code intelligence technology. Rather than just understanding the code in your current file, Cody can comprehend your entire codebase — including how different components relate, what APIs your code calls, and how changes in one place might affect others.

Cody is built on top of Sourcegraph's battle-tested code intelligence platform, which has been used by major enterprises for years to navigate and understand large codebases. This foundation gives Cody capabilities that are uniquely powerful for working in large, complex software projects where a single repository might contain millions of lines of code.

The AI can read across hundreds or thousands of files simultaneously, understanding the full architecture of an application. When you ask Cody to implement a new feature, it doesn't just generate generic code — it generates code that follows the patterns, conventions, and style already present in your specific codebase.

Cody supports multiple AI models, including Anthropic's Claude, and allows enterprise users to bring their own models. This flexibility lets organizations use the AI that best fits their security and capability requirements.

One of Cody's most valuable features is its precise context selection. When answering a question or generating code, Cody intelligently selects the most relevant parts of your codebase to include as context, rather than simply stuffing in random adjacent files. This makes responses more accurate and relevant to your specific situation.

The tool integrates with VS Code and JetBrains IDEs, as well as providing a web-based interface for those who prefer browser access. The chat interface allows for natural conversations about code, asking questions like "What does this function do?" or "How does the authentication system work?" and getting accurate, codebase-aware answers.

Cody also provides inline code suggestions as you type, learning from the context of surrounding code to predict what you're likely to write next. These autocomplete suggestions are informed by your entire codebase's patterns, not just local context.

For enterprise deployment, Cody can be connected to Sourcegraph's code search and intelligence platform, providing even deeper codebase understanding and security controls appropriate for large organizations.