Sourcegraph Cody
An AI coding assistant that leverages full codebase context for chat, completions, edits, and debugging.
What is Sourcegraph Cody?
Sourcegraph Cody is an AI-powered coding assistant introduced in June 2023 and generally available by December 14, 2023. It distinguishes itself by integrating deep retrieval‑augmented generation (RAG) architecture anchored in Sourcegraph's Code Graph and search infrastructure, enabling context‑aware chat, autocomplete, inline edits, prompts, and debugging across large codebases. It supports multiple IDEs including VS Code, JetBrains, Visual Studio (experimental), the web app, and CLI. Over time, Cody shifted focus purely to enterprises; free and pro tiers were sunset in mid‑2025, and individuals were redirected to the separate Amp product, while enterprise offerings remain embedded within Sourcegraph's Enterprise platform with full security, admin controls, and compliance.
What you can do with it
Code comprehension and navigation
Developers ask questions in natural language to understand how code works using context from across the codebase.
In-editor code generation and autocomplete
Cody predicts single- or multi-line code within a file based on surrounding context.
Automated refactoring or edits
Users highlight code and describe changes to get inline diffs generated and applied.
Debugging assistance
Cody identifies and suggests fixes for coding errors via chat or code‑action triggers.
Shared prompts for workflow automation
Teams create and share reusable prompts like 'explain function' or 'generate unit test' for consistency.
Key features
- Context‑aware code completions and suggestions powered by full codebase understanding
- Chat-based assistance that retrieves semantic context from local and remote repositories
- Inline code edits and refactoring via natural‑language prompts
- Customizable prompts for automating common development tasks (e.g. explanations, tests, documentation)
- Support for multiple IDEs and environments (VS Code, JetBrains, Visual Studio experimental, Web app, CLI)
- Context filtering to include or exclude specific repositories or symbols
- Enterprise-grade deployment with advanced security, audit logging, access control, and self‑hosting options
Screenshots
Inputs / Outputs
Strengths & Limitations
Strengths
Deep codebase awareness
Leverages Sourcegraph’s Code Graph and advanced search to provide AI assistance grounded in the full repository context rather than only open files.
Multi‑IDE and CLI support
Available across VS Code, JetBrains IDEs, Visual Studio (experimental), web application, and command‑line interface.
Enterprise‑grade features
Includes context filters, admin controls, audit logs, compliance (SOC 2, GDPR, CCPA), and cloud/self‑hosted deployment options.
Swappable LLM support
Supports a range of models from Anthropic, OpenAI, Google DeepMind, etc., with configurability in enterprise environments.
Limitations
No individual‑level access
Free and Pro tiers were discontinued mid‑2025, so individual developers can no longer access Cody.
Higher cost barrier
Access now requires Enterprise agreement – pricing can be substantial and lacks self‑serve plans.
Complex setup
Enterprise deployment may require integration with Sourcegraph infrastructure, increasing onboarding complexity.
Pricing & Plans
Model: Enterprise
Enterprise
Includes Cody AI credits and full Sourcegraph Enterprise features with security and admin controls
As of mid‑2025, only the Cody Enterprise tier remains available, structured around Sourcegraph Enterprise pricing (e.g. part of overall platform starting at approximately $16K/year or custom per‑user pricing). Free, Pro and mid‑tier Starter plans have been discontinued. Sourcegraph directs individual and small‑team developers to its separate Amp product.
Who it's for
Ideal for
Developers or engineering teams within medium to large organizations that already use or plan to use Sourcegraph Enterprise, and need AI assistance that understands entire codebases with robust security and administrative oversight.
Not ideal for
Individual developers or small teams seeking low‑cost, self‑serve AI coding tools—Sourcegraph directs this audience to its separate Amp product instead.
What users say
- powerful for large codebases
- preferred by enterprise teams
- context‑aware responses
- limited to enterprise use
Prompts & Results
›Explain this function in simple terms
Chat-based explanation of the function using codebase context to describe its purpose, inputs, outputs, and usage examples.
›Refactor this snippet for readability
Inline edited code with clearer variable names, simplified logic, and formatting consistent with project style.
›Generate unit tests for this module
Test code scaffold with multiple test cases covering expected behavior, edge cases, and error handling, using repository conventions.
›Find and fix the bug in this code
Cody identifies the likely error (e.g., off-by-one), suggests a corrected version, and adds comments explaining the fix using relevant surrounding code context.
FAQ
Is Cody still available to individual developers?+
No. The Free and Pro plans were discontinued as of June–July 2025. Sourcegraph now reserves Cody for Enterprise customers and directs individuals toward Amp.
Does Cody train on my code?+
No. For enterprise customers, Sourcegraph does not use customer code to train models; code is retrieved via search and fed to the LLM at runtime under strict permissions.
Which IDEs does Cody support?+
Cody supports VS Code, JetBrains IDEs, Visual Studio (experimental), Sourcegraph Web app, and a CLI interface.
What models does Cody use?+
Cody supports multiple LLMs from providers like Anthropic (Claude), OpenAI, Google DeepMind (Gemini), etc., and enterprise admins can configure which ones to allow.
Ratings & Reviews
No reviews yet — be the first to rate this tool.