Omni Dev AI Code Assistant
A VS Code pair-programming extension for the local or hosted Omni AI API and Ollama. It provides Copilot-style chat, inline ghost text, editor actions, and safe apply/preview flows.
Run the local backend
From the repository root:
npm install
npm run server
Ollama must be running with the configured model installed:
ollama serve
ollama pull llama3.2:3b
Install the extension locally
From this directory:
npm install
npm run compile
Then open this vscode-extension folder in VS Code and press F5 to launch an Extension Development Host. In the new VS Code window:
- Run
Omni AI: Check Local Connection from the Command Palette.
- Run
Omni AI: Open Chat or press Ctrl+Shift+O in an editor.
- Select code and use the
Omni AI editor context-menu commands for fixes, refactors, tests, explanations, and docstrings.
- Ghost-text completions are enabled by default.
Hosted login
Before chat, completions, or code actions can run, the extension requires an Omni AI account. Run Omni AI: Sign In; a browser window opens to the Omni AI login/signup page. After successful authentication, the browser redirects back to VS Code and the extension stores the session in VS Code Secret Storage. No password is stored by the extension.
Settings
omniDevAI.serverUrl: defaults to http://127.0.0.1:3001/api/pair-programmeromniDevAI.authUrl: hosted Omni AI login page, defaulting to http://localhost:5173/omniDevAI.model: defaults to ollama:llama3.2:3bomniDevAI.enableInlineCompletion: enables ghost-text completionsomniDevAI.completionDelay: debounce delay in milliseconds
