{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreiciv2hahfk5yxiacmcuasfnkjxzra7pz7tltqn6sskbmpb2txv6sq",
    "uri": "at://did:plc:5sgu76a53rz3n6unbykmovqy/app.bsky.feed.post/3mm33gjfegkv2"
  },
  "description": "Cursor is a fork of VS Code that reimagines the editor as an AI-native environment. While GitHub Copilot adds AI as a plugin to existing editors, Cursor was built from the ground up with AI as a first-class participant, not an afterthought.\n\n\nCore Features\n\n * Tab completion, predicts multi-line, multi-cursor edits based on recent changes and file context\n * Cmd+K inline editing, select code, describe what you want, Cursor rewrites it in place with a diff view\n * Chat (Cmd+L), context-aware chat",
  "path": "/engineering-glossary/cursor/",
  "publishedAt": "2026-05-17T19:20:49.000Z",
  "site": "https://sahilkapoor.com",
  "tags": [
    "Github Copilot",
    "Mcp Model Context Protocol",
    "System Prompt",
    "Openhands",
    "Windsurf",
    "Prompt Engineering"
  ],
  "textContent": "Cursor is a fork of VS Code that reimagines the editor as an AI-native environment. While GitHub Copilot adds AI as a plugin to existing editors, Cursor was built from the ground up with AI as a first-class participant, not an afterthought.\n\n## Core Features\n\n  * **Tab completion** , predicts multi-line, multi-cursor edits based on recent changes and file context\n  * **Cmd+K inline editing** , select code, describe what you want, Cursor rewrites it in place with a diff view\n  * **Chat (Cmd+L)** , context-aware chat that understands your whole codebase, not just the open file\n  * **Agent mode (Cmd+Shift+I)** , the AI runs autonomously: reads files, writes code, runs terminal commands, iterates on errors\n  * **Codebase indexing** , Cursor embeds your entire repo for semantic search, so it can find relevant context across thousands of files\n\n\n\n## Model Flexibility\n\nCursor supports multiple underlying models: Claude Sonnet/Opus, GPT-4o, Gemini, and smaller fast models for tab completion. You can bring your own API key or use Cursor's subscription (which includes usage credits). The choice of model affects quality, speed, and cost per interaction.\n\n## How It Differs from Copilot\n\nGithub Copilot is primarily a suggestion engine, it autocompletes as you type. Cursor's agent mode can autonomously complete multi-step tasks: \"refactor this module to use async/await, update all the tests, and fix the type errors.\" This shifts the paradigm from _suggestion_ to _delegation_.\n\n## MCP Support\n\nCursor supports Mcp Model Context Protocol servers, allowing the AI to interact with external tools, running SQL queries, fetching docs, reading your CI results, directly within the editor context.\n\n## Cursor Rules\n\nYou can define project-level rules (in `.cursor/rules`) that tell the AI how to behave: coding conventions, architectural patterns, libraries to prefer, things to avoid. These feed into every interaction as additional System Prompt content.\n\n## When to Use Cursor vs Other Tools\n\n  * For solo development where you want the AI as a coding partner: Cursor\n  * For teams already on VS Code/JetBrains wanting minimal disruption: Github Copilot\n  * For autonomous agents running unattended: Openhands\n  * For lightweight in-browser edits: Windsurf\n\n\n\n## Related Terms\n\n  * Github Copilot, AI coding assistant; narrower feature set, broader IDE support\n  * Windsurf, competing AI IDE from Codeium with similar feature set\n  * Openhands, open-source autonomous coding agent\n  * Mcp Model Context Protocol, protocol Cursor uses for external tool access\n  * Prompt Engineering, skill that makes Cursor's chat interactions more effective\n\n",
  "title": "Cursor",
  "updatedAt": "2026-05-18T20:04:11.251Z"
}