OpenCode Zen Provider

VS Code extension project for using curated OpenCode Zen models in Copilot Chat with your own OpenCode Zen API key.

Requirements

• VS Code 1.104.0 or later
• GitHub Copilot extension installed and active
• An OpenCode Zen API key from https://opencode.ai/auth

Installation

From Source

1. Clone this repository.
2. Run bun install --ignore-scripts && bun run compile.
3. Press F5 in VS Code to launch the Extension Development Host.

From VSIX

1. Run bun install --ignore-scripts && bun run package:vsix.
2. Install the generated .vsix file via Install from VSIX... in the Extensions view.

Setup

1. Open Copilot Chat and open the model picker.
2. Add or configure OpenCode Zen.
3. Enter your OpenCode Zen API key when prompted.
4. If needed, run OpenCode Zen: Manage OpenCode Zen API Key from the Command Palette.
5. Select OpenCode Zen in Copilot Chat and choose a model.

Supported Models

The extension bundles a static ZEN_MODEL_CATALOG (in src/model-catalog.ts) that mirrors the current OpenCode Zen API model lineup, including:

• DeepSeek: V4 Pro, V4 Flash, V4 Flash Free
• Gemini: 3 Flash, 3.5 Flash Lite, 3.6 Flash, 3.7 Flash
• GPT: 5.6 Luna / Sol / Terra
• GLM: 5.2
• Grok: Build 0.1, 4.5, 4.6
• Kimi: K3
• MiniMax: M3
• Muse: Spark 1.2, Spark 1.2 Contributor Free
• X Preview F Free (preview)
• Free models: Big Pickle, Hy3 Free, MiMo V2.5 Free, Nemotron 3 Ultra Free, Nemotron 3.5 Lightning Free, Laguna S 2.1 Free

For the full table of capabilities (context window, vision, tools, thinking, API format), see docs/models.md.

Architecture

• Open the interactive architecture diagram — Request path, key storage, model routing, and image-analysis fallback

Documentation

• Supported Models — Full model list, capabilities, context window, and model quirks

Current V1 Scope

• Provider registration, API key management, static model catalog, and core chat-provider wiring are included.
• The custom image-analysis tool is available via opencode_zen_analyze_image and automatic image fallback for non-vision models.
• Non-vision models either switch to a vision-capable fallback model or delegate to the image-analysis tool when images are attached.

Development

bun install --ignore-scripts
bun run compile
bun run lint
bun run test -- --runInBand

Press F5 in VS Code to launch the Extension Development Host.

Available Scripts

• bun run compile - TypeScript compile
• bun run watch - TypeScript watch mode
• bun run test - Run Jest
• bun run lint - Run ESLint
• bun run lint:fix - Auto-fix ESLint issues
• bun run format - Run Prettier
• bun run package:vsix - Build a VSIX package

Marketplace Packaging

Build the VSIX package:

bun run package:vsix

Publishing is automated by .github/workflows/publish.yml: pushing a v* tag publishes the VSIX to the VS Code Marketplace (repository secret VSCE_PAT). This extension only registers a GitHub Copilot Chat model provider (languageModelChatProviders), and Copilot Chat does not exist in the editors Open VSX serves, so it is not published there.

To publish a VSIX manually:

vsce publish --packagePath opencode-zen-provider-<version>.vsix --allow-missing-repository   # VSCE_PAT

Privacy

• Your API key is stored securely in VS Code SecretStorage.
• The Zen provider base URL is https://opencode.ai/zen/v1.