Neuron by SADATA Provider

VS Code extension that connects models from https://ask.sadata.id into the VS Code Chat model picker through the stable registerLanguageModelChatProvider API.

Features

• Fixed backend endpoint: https://ask.sadata.id
• API key stored securely via VS Code SecretStorage
• Connected-state UI optimization:
  • if key is already saved, key input form is hidden
  • actions shown are Refresh and Disconnect
• Model loading from /api/v1/models with cache and refresh flow
• Runtime tool-calling support probe through chat completions (not inferred from model metadata alone)
• Streaming chat response handling
• Status bar analytics and quick actions
• Output channel for diagnostics: Neuron

Requirements

• VS Code 1.104.0 or newer

Development

1. Open this folder in VS Code.
2. Press F5 to run Extension Development Host.
3. Open Neuron in the Activity Bar.
4. Set API key from ask.sadata.id.
5. Use VS Code Chat model picker to select available models.

Build and Install

npm run lint
npm run package
code --install-extension dist/sadata-neuron.vsix

Architecture

src/extension.js  - activation, command registration, status bar
src/provider.js   - LanguageModelChatProvider implementation
src/api.js        - HTTP and streaming integration
src/config.js     - constants and secret helpers
src/webview.js    - connection webview controller
media/view.html   - webview markup
media/view.css    - webview styling
media/view.js     - webview behavior

Versioning and Changelog

CHANGELOG.md is the release reference.

SemVer policy:

• Patch (x.y.Z): bug fix, UI tweak, non-breaking maintenance
• Minor (x.Y.z): new feature, backward-compatible change
• Major (X.y.z): breaking behavior or compatibility change

For every change that ships:

1. Update version in package.json
2. Keep package-lock.json version in sync
3. Add a clear entry in CHANGELOG.md

Tools and Skills Note

Any tool or skill execution still happens on the backend side. The extension only surfaces model responses and related metadata in VS Code.