What it is
Cursor describes itself as a coding agent for building ambitious software. It is offered as a desktop app (Cursor Desktop) and a CLI (Cursor Agent). Agents can plan and build features, and cloud agents work autonomously and in parallel, using their own computers to build, test and demo features. Cursor also runs in Slack and reviews PRs in GitHub.
Who it's for
- Developers who want to hand off coding tasks to AI agents while they focus on decisions
- Teams that want AI assistance in the terminal, Slack and GitHub
- Enterprises looking for AI-assisted development at scale
Requirements
Requirements
- A macOS machine for the Download for macOS option shown on the homepage
Setup
Download the desktop app
The homepage offers a 'Download for macOS' button and a 'Get started' button.
Install from the command line
The homepage shows this command under 'Cursor runs in your terminal'. Once installed, the CLI demo starts an agent with the
agentcommand and shows '/ commands · @ files · ! shell' hints.bashcurl https://cursor.com/install -fsS | bash
Examples
Generate a landing page
Promptmake a landing page based on attached docs explaining what we doExpected output: A prompt from the Cursor Desktop demo, which builds a landing page from attached documents.
Extend a bioinformatics tool via the CLI
PromptAdd affine gap alignment. We only have basic NW alignment but I need Gotoh's algorithm for better handling of indels. Also fix the FASTA parser.Expected output: A prompt from the Cursor CLI demo, run in a repository called 'Bioinformatics Tools'.
Plan a feature with an agent
Promptlet's build a mission control interface, similar to the expose-style window manager on macOSExpected output: In the demo this produces a plan (feature-prd.md) and a task list covering the new view component and the app manager.
Delegate a task from Slack
Prompt@cursor can you take a stab?Expected output: In the Slack demo, mentioning @cursor in a thread leads to a PR. The reply offers View PR, Open in Cursor and Open in Web.
Pros & cons
Pros
- Pro:Supports models from OpenAI, Anthropic, Gemini, SpaceXAI and Cursor, so you can pick one per task
- Pro:Cloud agents run in parallel on long tasks, for hours or days
- Pro:Works in the terminal, Slack and GitHub as well as the desktop editor
- Pro:Offers a range of autonomy, from targeted Cmd+K edits to fully agentic runs
Cons
- Con:The homepage offers only a macOS download; other platforms are not mentioned
- Con:The source is a marketing page with no pricing, system requirements or detailed setup documentation