Skip to content

feat: auto-sync models from CommandCode API at startup #1

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
augustoolucas wants to merge 3 commits into
base: main
Choose a base branch
from
Open

feat: auto-sync models from CommandCode API at startup #1

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
095ed31
feat: auto-sync models from CommandCode API at startup
augustoolucas Jun 4, 2026
33a3703
feat: harden auto-sync (in-memory, opt-out, fail-loud load)
augustoolucas Jun 5, 2026
cf25a8b
fix: path-aware mocks, unexport mergeModels, scoped readFileSync
augustoolucas Jun 6, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
36 changes: 29 additions & 7 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,7 +46,7 @@ If you prefer to configure manually, add this to your `opencode.json`:
}
```

The plugin auto-registers models from [`models.json`](./models.json) at startup. You only need the `provider.commandcode` block — no need to list individual models.
The plugin auto-registers models from [`models.json`](./models.json) at startup, with specs automatically synced from the CommandCode API. You only need the `provider.commandcode` block — no need to list individual models.

### Environment Variable

Expand All @@ -56,17 +56,34 @@ Set `COMMANDCODE_API_KEY` instead of using `/connect`:
COMMANDCODE_API_KEY=your-key opencode
```

### Disabling Auto-Sync

By default, the plugin syncs model specs from the CommandCode API on every opencode startup. The sync is best-effort and silent — if the API is unreachable, returns invalid JSON, or the `COMMANDCODE_API_KEY` is invalid, the plugin falls back to the bundled [`models.json`](./models.json).

To disable the sync, create `~/.config/opencode/commandcode-go-opencode-provider.json`:

```json
{
"disableModelSync": true
}
```

When this file exists with `disableModelSync: true`, the plugin skips the API fetch entirely and uses the bundled `models.json` as-is.

**Note:** Bundled `models.json` is updated by maintainer releases, not at runtime. Run `bun run sync` to refresh from the latest Command Code API.

## Available Models

| Model ID | Name | Tier | Reasoning | Context |
|---|---|---|---|---|
| `claude-haiku-4-5-20251001` | Claude Haiku 4.5 | premium | no | 200K |
| `claude-opus-4-7` | Claude Opus 4.7 | premium | yes | 1M |
| `claude-opus-4-8` | Claude Opus 4.8 | premium | yes | 1M |
| `claude-sonnet-4-6` | Claude Sonnet 4.6 | premium | yes | 1M |
| `gpt-5.3-codex` | GPT-5.3 Codex | premium | yes | 400K |
| `gpt-5.4` | GPT-5.4 | premium | yes | 400K |
| `gpt-5.4-mini` | GPT-5.4 Mini | premium | yes | 400K |
| `gpt-5.5` | GPT-5.5 | premium | yes | 256K |
| `gpt-5.5` | GPT-5.5 | premium | yes | 200K |
| `deepseek/deepseek-v4-flash` | DeepSeek V4 Flash | open-source | yes | 1M |
| `deepseek/deepseek-v4-pro` | DeepSeek V4 Pro | open-source | yes | 1M |
| `google/gemini-3.1-flash-lite` | Gemini 3.1 Flash Lite | open-source | yes | 1M |
Expand All @@ -76,13 +93,18 @@ COMMANDCODE_API_KEY=your-key opencode
| `moonshotai/Kimi-K2.5` | Kimi K2.5 | open-source | no | 256K |
| `moonshotai/Kimi-K2.6` | Kimi K2.6 | open-source | no | 256K |
| `MiniMaxAI/MiniMax-M2.5` | MiniMax M2.5 | open-source | no | 200K |
| `MiniMaxAI/MiniMax-M2.7` | MiniMax M2.7 | open-source | no | 1M |
| `Qwen/Qwen3.6-Max-Preview` | Qwen 3.6 Max Preview | open-source | yes | 1M |
| `Qwen/Qwen3.6-Plus` | Qwen 3.6 Plus | open-source | yes | 1M |
| `MiniMaxAI/MiniMax-M2.7` | MiniMax M2.7 | open-source | no | 200K |
| `MiniMaxAI/MiniMax-M3` | MiniMax M3 | open-source | no | 1M |
| `Qwen/Qwen3.6-Max-Preview` | Qwen 3.6 Max Preview | open-source | yes | 200K |
| `Qwen/Qwen3.6-Plus` | Qwen 3.6 Plus | open-source | yes | 200K |
| `Qwen/Qwen3.7-Max` | Qwen 3.7 Max | open-source | yes | 1M |
| `Qwen/Qwen3.7-Plus` | Qwen 3.7 Plus | open-source | yes | 1M |
| `stepfun/Step-3.5-Flash` | Step 3.5 Flash | open-source | yes | 1M |
| `stepfun/Step-3.7-Flash` | Step 3.7 Flash | open-source | yes | 256K |
| `xiaomi/mimo-v2.5` | Xiaomi MiMo V2.5 | open-source | no | 1M |
| `xiaomi/mimo-v2.5-pro` | Xiaomi MiMo V2.5 Pro | open-source | no | 1M |

Full model list is maintained in [`models.json`](./models.json). Run `bun run sync` to refresh from the latest Command Code CLI release on npm.
Models are automatically synced from the CommandCode API at startup. Context sizes and new models are updated on each restart. The local [`models.json`](./models.json) serves as an offline fallback.

## Development

Expand All @@ -109,7 +131,7 @@ For local testing, create `opencode.local.json` (gitignored) with `file://` path

Run `opencode --config opencode.local.json` to test with your local build.

### Sync Models
### Sync Models (Manual)

```bash
bun run sync # update models.json from Command Code
Expand Down
Loading