Sourcegraph Cody vs Cursor
Cursor is an “AI Code Editor.” Unlike other solutions which are commonly IDE extensions and plugins, Cursor itself is a fork of VS Code. Cursor features autocomplete (called "Tab"), chat, Agent, Composer, and Command+K for quickly editing code. Cursor fetches context based off file and folders on your local machine, but can't use context from remote repositories or the broader codebase.
The main downside of Cursor being a fork of VS Code is that it is not available to use with other IDEs (such as the JetBrains family). It also doesn't maintain complete feature parity with VS Code. For users with large codebases - particularly in the enterprise - Cursor’s lack of support for remote codebase context may mean results aren’t as accurate.
Cursor is a good choice for users working in smaller codebases who want an AI-centric IDE, and are willing to migrate from their preferred IDE to Cursor. It offers unique functionality and has local codebase context features. However, Cody is a better choice for users who want to use their existing IDE, or for enterprises that want to make their AI code assistant accessible to many of their developers who may be working in various IDEs. Cody also offers more functionality for enterprise developers, including the ability to reference remote codebase context in chat and flexibility in LLM deployment: for example, securely connecting to LLMs via Amazon Bedrock. Lastly, Cody offers more generous usage for premium models such as Claude 3.7 Sonnet and GPT-4o on the Pro tier (whereas usage is limited for premium models on Cursor's plans before switching to “slow requests” or additional usage based pricing).
