What is Cursor AI? Features, pricing, and Geek Hub setup
Cursor is a Visual Studio Code–based editor with LLMs baked into the editing loop. You keep the familiar VS Code UI while getting multi-line Tab completions, chat/agent flows, and multi-file change modes. Many teams now treat it as their primary AI coding IDE.
What is Cursor?
Cursor started as a VS Code fork: same shortcuts, extensions, and workspaces, plus native AI. Instead of pasting code into an external chat, the model sees repo context (open files, diffs, project rules) and proposes edits in-editor.
It pairs a fast Tab experience for next-edit prediction with cloud agents that can touch many files, run the terminal, and open PRs depending on your plan.
Key features
What usually matters when comparing Cursor to peers:
- Tab / inline: multi-line edits without leaving the file
- Agent / Composer: multi-file tasks with planned diffs
- Codebase context and project rules for style alignment
- Configurable models, including OpenAI-compatible base URLs
- Privacy Mode on team plans to reduce code retention
Who it's for
Best if you already live in VS Code and want AI without switching IDEs. Strong on bounded refactors, tests, bugs, and mid-size features. For huge migrations, split the work: Cursor speeds the first draft; human review remains the real bottleneck.
Pricing (indicative)
Cursor offers a free evaluation tier, Pro for individuals (~$20/mo with model credit), and Teams/Enterprise with SSO and privacy controls. Hosted model usage can add cost—many teams use a gateway like Geek Hub.
Using Cursor with Geek Hub
Geek Hub exposes an OpenAI-compatible API. In Cursor, add a custom OpenAI-compatible provider and set the base URL to Geek Hub—then switch Claude, GPT, or Gemini by model id alone.
- Create an account at app.geekhub.mx and mint an API key
- Cursor → Settings → Models → add an OpenAI-compatible provider
- Base URL: https://api.geekhub.mx/v1
- Paste your Geek Hub key and pick a catalog model id
- Try Tab and Agent; watch spend in the Geek Hub dashboard
OPENAI_BASE_URL=https://api.geekhub.mx/v1
OPENAI_API_KEY=gh_...
# model example: openai/gpt-5.4-miniGet an API keyCursor FAQ
- Does Cursor work with Geek Hub?
- Yes, when you set an OpenAI-compatible endpoint. Geek Hub is the single base URL to multiple providers.
- Will a gateway slow me down?
- Typical overhead is a few ms. You gain one key, failover options, and clearer per-model cost.
- Can I use Claude and GPT in one project?
- Yes—change the model id in Cursor; Geek Hub routes to the right provider.