Does Sourcegraph Cody require Sourcegraph Enterprise?

No. Sourcegraph Cody works standalone with local file context. Sourcegraph Enterprise unlocks cross-repo graph context and SSO.

Which models does Sourcegraph Cody support?

Sourcegraph Cody supports Claude, GPT, Gemini, and local open models. Pro and Enterprise users can switch per request.

How does Sourcegraph Cody compare to GitHub Copilot?

Sourcegraph Cody fetches context from your entire org's repos via Sourcegraph search. Copilot is scoped to the files you have open.

Can I self-host Sourcegraph Cody?

Yes, with Sourcegraph Enterprise you can self-host the entire Cody stack, including model proxying, behind your own firewall.

Sourcegraph Cody

AI Assistant That Knows Your Entire Codebase

Freemium

Code AssistantPair ProgrammingIDE

Visit Sourcegraph Cody

What is Sourcegraph Cody?

Sourcegraph Cody is an AI coding assistant that leverages Sourcegraph's code intelligence platform to understand your entire codebase at scale. Sourcegraph Cody fetches context from across your repositories, dependencies, and documentation before answering questions or generating code. Sourcegraph Cody supports Claude, GPT, Gemini, and open models, and runs in VS Code, JetBrains, and Neovim. Sourcegraph Cody excels where other tools lose context: sprawling monorepos with millions of lines, multi-repo microservice architectures, and large teams where tribal knowledge lives across many projects. Sourcegraph Cody's @-mention syntax lets you attach files, symbols, or entire repos to a prompt, giving it the precise context needed for accurate answers.

Key Features of Sourcegraph Cody

Codebase-Wide Context

Sourcegraph Cody uses Sourcegraph's search and index to pull context from any repo in your organization. It answers questions about code you haven't opened, unlike editor-only tools.

@-Mention Syntax

Prefix your prompts with @file, @symbol, or @repo to attach exactly the context Sourcegraph Cody needs. This eliminates copy-paste and ensures precise, grounded answers.

Multi-Model Support

Sourcegraph Cody lets you pick between Claude, GPT, Gemini, and local models per request. Use a powerful model for architecture questions and a cheap one for quick completions.

Inline Edits

Select code, describe a change, and Sourcegraph Cody rewrites it in place with a diff you can accept or reject. This works across multiple selections in the same file.

Commands and Recipes

Sourcegraph Cody ships built-in commands like /explain, /test, /doc, and /smell. Create custom recipes your team shares so everyone runs the same AI workflows.

Enterprise Graph Context

With Sourcegraph Enterprise, Sourcegraph Cody reasons over precise code intelligence: type definitions, references, and dependency graphs across every indexed repository.

Sourcegraph Cody Pricing Plans

Free

Sourcegraph Cody for individuals with 500 autocompletes and 200 chat messages per month using default models. Good for evaluating Sourcegraph Cody.

Pro

$9/month

Unlimited Sourcegraph Cody autocompletes, 1000 chat messages, and access to frontier models. Designed for professional developers on personal projects.

Enterprise

$19/user/month

Sourcegraph Cody with full codebase context, SSO, unlimited chat, custom recipes, and admin controls. Ties into Sourcegraph Enterprise search.

Best Use Cases for Sourcegraph Cody

Debugging Across Microservices

Target user:: Engineer tracing a bug through five repos
Pain point:: Each repo needs a separate clone, search, and context switch, and the bug crosses boundaries
Solution:: Ask Sourcegraph Cody to search all five repos for the error string. It returns the exact files and trace path across services in seconds.

Writing Cross-Repo Refactors

Target user:: Staff engineer updating a shared API contract
Pain point:: You need to find every consumer and update call sites you didn't write
Solution:: Use @repo to give Sourcegraph Cody the API definition, then ask for all call sites. It generates patches for every consumer in one pass.

Answering Architecture Questions

Target user:: Tech lead evaluating a refactor proposal
Pain point:: You need to know how many services depend on Module X before you can approve the change
Solution:: Sourcegraph Cody queries the code graph and counts reverse dependencies instantly, citing the exact import paths for each dependent service.

Onboarding to a Large Monorepo

Target user:: New engineer at a company with a 2M-line repo
Pain point:: Standard AI tools timeout or hallucinate because the codebase exceeds their context window
Solution:: Sourcegraph Cody fetches only the relevant slices through Sourcegraph search, giving accurate answers without needing to load the entire repo.

How to Use Sourcegraph Cody — Step by Step

1
Install the Extension
Install Sourcegraph Cody from the VS Code or JetBrains marketplace. Create a free Sourcegraph account to activate the assistant.
2
Connect Your Repos
Point Sourcegraph Cody to your Sourcegraph instance or let it index local files. Enterprise users get organization-wide code intelligence automatically.
3
Ask a Codebase Question
Open the Sourcegraph Cody chat and type a question like 'how does user auth work?' Use @repo to include files outside your current editor.
4
Apply an Inline Edit
Select code, press the Sourcegraph Cody edit hotkey, and describe the change. Review the diff and accept or reject each hunk individually.
5
Use Built-In Commands
Run /explain on a complex function, /test on a module, or /doc to generate documentation. Customize commands in your Sourcegraph Cody settings.

Sourcegraph Cody vs Alternatives

Need context across many large repos?

Pick: Sourcegraph CodySourcegraph Cody's code search and graph give it context that editor-local tools simply cannot access without manual copy-paste.

Want a full AI-first editor?

Pick: CursorSourcegraph Cody is a plugin, not an editor. For deep agentic editing with a custom IDE, Cursor provides a more integrated experience.

Just need fast inline completions?

Pick: GitHub Copilot or WindsurfFor pure speed and polish on single-file completion, Copilot and Windsurf are simpler. Sourcegraph Cody shines on multi-repo reasoning.