PawanOsman/OpenCursor

Open-source Cursor-like AI coding agent for VS Code — agentic chat, multi-provider LLMs (OpenAI, Ollama, llama.cpp), semantic search, and MCP support

What it solves

OpenCursor provides an open-source, local-first AI coding agent for VS Code, allowing developers to use AI assistance without relying on cloud services or sending their code to external servers.

How it works

It integrates directly into VS Code as an extension. Users can choose between local models (via llama.cpp or Ollama), API keys (OpenAI, Anthropic, Gemini, OpenRouter), or OAuth sign-ins. The agent uses a local ONNX MiniLM model for semantic codebase search, meaning it indexes and searches code by meaning rather than keywords entirely on-device. It employs a set of 25 tools to read, write, and edit files, run shell commands, and perform web searches.

Who it’s for

Developers who want a powerful AI coding assistant in VS Code but require high privacy, offline capabilities, or the flexibility to use their own model providers and subscriptions.

Highlights

  • 100% Local AI: Built-in support for llama.cpp and Ollama, enabling fully offline coding.
  • Semantic Search: Local codebase indexing and search powered by on-device embeddings.
  • Flexible Providers: Supports OAuth sign-ins, API keys, and custom OpenAI-compatible endpoints.
  • Agentic Tooling: Includes 25 tools for file manipulation, shell access, and MCP (Model Context Protocol) support.
  • Safety Controls: Per-action approval policies with risk heuristics to prevent dangerous commands.
  • Multimodal Support: Ability to process images and PDFs directly into the chat context.

Related

  • Project
    StartupHakk/OpenMonoAgent.ai

    A local-first, open-source coding agent that runs entirely on your hardware using .NET and llama.cpp to provide private, zero-cost AI development.

  • Project
    Nano-Collective/nanocoder

    An open-source, local-first coding agent for the terminal that allows developers to use their own AI models via Ollama or various APIs while keeping code private.

  • Project
    ggml-org/llama.vscode

    A VS Code extension that enables local LLM-assisted text completion, chat, and agentic coding by integrating with llama.cpp.

  • Project
    ochyai/vibe-local

    A free, offline AI coding environment for Mac and Linux that integrates Ollama and OpenCode to provide local LLM-powered agentic coding without API costs.

  • Dispatch
    Ollama and Continue: Setting Up an Open-Source AI Code Assistant

    Ollama and Continue enable developers to build a fully open-source, local AI coding assistant within VS Code and JetBrains using models like Codestral, Llama 3, and DeepSeek Coder.