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 is 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
Overview of OpenCursor
What it is
OpenCursor is an open-source AI coding agent for VS Code. It brings conversation, code exploration, file edits, terminal commands, and change review into the editor. Users can connect a supported account, use their own API keys, or run a local model with Ollama or llama.cpp. The agent works with multiple LLM providers and allows choosing the model and permissions for each task.
Key features
- Workspace-aware chat: mention files, folders, selections, documentation, terminal output, and Git context; keep separate conversations in tabs with saved drafts and searchable history.
- Rich messages: write Markdown, paste images, preview attachments, and use highlighted code blocks with automatic language detection; supports RTL and LTR content.
- Visible progress: follow collapsible work activity, tool calls, task lists, and subagents; completed work keeps its summary easy to find.
- Tools for implementation: read and edit files, search code, run terminal commands, inspect language symbols, and verify browser behavior.
- Change review: inspect changed files and review individual hunks in the editor; use Keep All or Undo All when ready.
- Multiple modes: Agent, Ask, Plan, Project, and Multitask; Ask is read-only, Plan can save a plan without editing project code, Agent carries out changes under configured permissions.
- Semantic search: combines embeddings with lexical matches to find code even when the question uses different words; index updates incrementally and accounts for unsaved editor content.
Use cases
- Coding assistance with workspace-aware chat and file editing.
- Running terminal commands and reviewing changes within the editor.
- Working offline with local models (Ollama, llama.cpp) and local embeddings.
- Indexing documentation with defined scope and referencing it via @Docs.
- Delegating independent work to subagents with optional token budgets.
- Connecting external tools and resources through MCP servers.
- Using isolated Git worktrees or Docker command execution for separate tasks.
Getting started
See the repository README for installation and quick-start instructions.
License and pricing
MIT license. Open-source and free to use.
Based on the project's public README and repository metadata.