An open-source AI coding agent that runs in your terminal, with companion desktop, web, editor, and chat interfaces.
PawanOsman/OpenCursor
Open-source AI coding agent for VS Code that works with your own models, including hosted APIs, accounts, and local Ollama or llama.cpp.
Open-source alternative to:Cursor
GitHub Copilot
- Stars
- 6,038
- Forks
- 1,028
- Open issues
- 6
- Last commit
- 26 days ago
What is it?
OpenCursor is a Cursor-style chat and coding agent that runs inside VS Code. It can read and edit files, search code, run terminal commands, and show each file change so you can keep or undo it. Tasks run in modes such as Agent, Ask (read-only), and Plan, and permission settings control which actions need approval.
The extension is model-agnostic. You can connect accounts such as GitHub Copilot or Claude Code, add API providers like OpenAI, Anthropic, Gemini, or OpenRouter, point it at any OpenAI- or Anthropic-compatible endpoint, or run local models through Ollama or llama.cpp. Multiple credentials for one provider can be balanced or failed over.
Context comes from a local semantic index of the workspace that combines embeddings with text matching, plus optional documentation sources that are indexed with a defined page scope. It also supports MCP servers, hooks, reusable skills, Git worktrees, and Docker-based command execution.
Key features
- 01Agent, Ask, and Plan modes with configurable permissions for file edits and commands
- 02Connects to hosted APIs, subscription accounts, custom OpenAI- or Anthropic-compatible endpoints, and local Ollama or llama.cpp models
- 03Semantic and lexical code search with an incremental local index
- 04Per-hunk change review with Keep All and Undo All
- 05MCP server support, hooks, reusable skills, and subagents
Use cases
- Running an AI coding assistant fully offline with a local GGUF model and local embeddings
- Using existing API keys or subscription accounts inside VS Code without switching editors
- Asking questions about a codebase in read-only Ask mode before making changes
- Indexing a specific library's documentation so the agent can reference it with @Docs
Quick start
- 1
Install OpenCursor Agent from the VS Code Marketplace
- 2
Connect a model under Providers in the OpenCursor Settings, or configure Ollama or llama.cpp
- 3
Enable a model in Models, select it in the composer, and describe your task
- 4
Review tool activity and file changes, then keep or undo the edits
Things to know
- Requires VS Code 1.96 or newer.
- Some features download native runtimes or model files on first use, so offline use needs preparation.
- Hosted models need your own accounts or API keys, and availability depends on the provider.
Topics
FAQ
- What is OpenCursor?
- Open-source AI coding agent for VS Code that works with your own models, including hosted APIs, accounts, and local Ollama or llama.cpp. OpenCursor is a Cursor-style chat and coding agent that runs inside VS Code. It can read and edit files, search code, run terminal commands, and show each file change so you can keep or undo it. Tasks run in modes such as Agent, Ask (read-only), and Plan, and permission settings control which actions need approval.
- Is OpenCursor free to use? What license does it have?
- OpenCursor is open source under the MIT license. Check the license terms before using it in commercial products.
- Is OpenCursor still maintained?
- Actively maintained. The last commit to OpenCursor was on Sep 25, 2026.
- What can I use OpenCursor for?
- Running an AI coding assistant fully offline with a local GGUF model and local embeddings Using existing API keys or subscription accounts inside VS Code without switching editors Asking questions about a codebase in read-only Ask mode before making changes Indexing a specific library's documentation so the agent can reference it with @Docs
- How do I install OpenCursor?
- Install OpenCursor Agent from the VS Code Marketplace Connect a model under Providers in the OpenCursor Settings, or configure Ollama or llama.cpp Enable a model in Models, select it in the composer, and describe your task Review tool activity and file changes, then keep or undo the edits
- Where can I get OpenCursor?
- The source code and installation instructions are on GitHub: https://github.com/PawanOsman/OpenCursor
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open source AI coding agent that runs in VS Code, JetBrains IDEs, or the terminal, with access to 500+ models.
Open-weight 4B and 1.7B general-purpose language models from XHToken, aimed at on-device use and agentic tool-calling workflows.
A VS Code extension that turns plain .chat.md files into editable AI chat sessions with MCP tool support across LLM providers.
Agent-native memory infrastructure that turns LLM agent conversations and execution into persistent, structured memory for Python and TypeScript apps.
An agent framework that connects digital humans (2.5D, 3D, mobile, PC, web) and LLMs (OpenAI-compatible, DeepSeek) to business systems.