Skip to content

README.md

Latest commit

 

History

History
227 lines (176 loc) · 11.7 KB

README.md

File metadata and controls

227 lines (176 loc) · 11.7 KB
Raw

opencode-commandcode-provider

Command Code API provider for opencode. Use Claude, GPT, Gemini, DeepSeek, Qwen, Kimi, GLM, MiniMax, Step, and other models through a single API key.

This is a community fork — install directly from GitHub:

Quick Start

1. Install

opencode plugin https://github.com/yoni13/opencode-commandcode-provider

This installs the provider and registers all available models automatically.

2. Connect

Run /connect in opencode, search for Command Code, and enter your API key:

/connect

Multiple keys can be entered as a comma-separated list:

key1,key2,key3

When a key reports insufficient credits, the provider switches to the next key. Key values are never included in status messages or logs.

3. Select a model

Run /models to pick from available models:

/models

Manual Configuration

If you prefer to configure manually, add this to your opencode.json:

{
  "plugin": ["https://github.com/yoni13/opencode-commandcode-provider/server"],
  "provider": {
    "commandcode": {
      "npm": "https://github.com/yoni13/opencode-commandcode-provider",
      "name": "Command Code",
      "env": ["COMMANDCODE_API_KEY"],
      "options": {
        "excludePremiumModels": true
      }
    }
  },
  "model": "commandcode/deepseek-v4-flash"
}

The plugin auto-registers models from models.json at startup. You only need the provider.commandcode block — no need to list individual models.

excludePremiumModels: true removes every model marked as premium from the opencode model picker. It defaults to false.

The plugin checks this fork's latest models.json at startup and falls back to its bundled catalog if the request fails, times out, is invalid, or is older than the bundled catalog. Set autoUpdateModels: false under provider.commandcode.options to disable this check. A scheduled GitHub workflow refreshes the repository catalog daily from the latest Command Code CLI and pricing page.

Environment Variable

Set COMMANDCODE_API_KEY instead of using /connect:

COMMANDCODE_API_KEY=your-key opencode

The environment variable also accepts comma-separated keys:

COMMANDCODE_API_KEY=key1,key2,key3 opencode

Reasoning Effort

Some Command Code models support explicit reasoning effort. When calling the provider through the AI SDK, pass the setting under the commandcode provider options:

import { streamText } from "ai"
import { createCommandCode } from "commandcode-go-opencode-provider"

const commandcode = createCommandCode()

await streamText({
  model: commandcode.languageModel("claude-sonnet-4-6"),
  prompt: "Implement the next task.",
  providerOptions: {
    commandcode: {
      reasoningEffort: "high",
    },
  },
})

The provider sends this as reasoning_effort to Command Code. Supported values are low, medium, high, xhigh, and max, depending on the selected model. The generated models.json records per-model support in reasoning_efforts and exposes them to opencode as model variants, so the opencode variant picker can select the effort when the installed opencode build supports variants.

Vision Inputs

Models that Command Code publishes with image input support are exposed to opencode with attachment: true and modalities.input: ["text", "image"]. Image inputs from the AI SDK are forwarded as Command Code image content, including URL, data URL, base64 string, and Uint8Array file parts.

Available Models

Model ID Name Tier Reasoning Context
claude-fable-5 [Premium] Claude Fable 5 premium yes 1M
claude-haiku-4-5-20251001 [Premium] Claude Haiku 4.5 premium no 200K
claude-opus-4-7 [Premium] Claude Opus 4.7 premium yes 1M
claude-opus-4-8 [Premium] Claude Opus 4.8 premium yes 1M
claude-opus-5 [Premium] Claude Opus 5 premium yes 1M
claude-sonnet-4-6 [Premium] Claude Sonnet 4.6 premium yes 1M
claude-sonnet-5 [Premium] Claude Sonnet 5 premium yes 1M
google/gemini-3.1-flash-lite [Premium] Gemini 3.1 Flash Lite premium yes 1M
google/gemini-3.5-flash [Premium] Gemini 3.5 Flash premium yes 1M
google/gemini-3.5-flash-lite [Premium] Gemini 3.5 Flash Lite premium yes 1M
google/gemini-3.6-flash [Premium] Gemini 3.6 Flash premium yes 1M
google/gemini-3.7-flash [Premium] Gemini 3.7 Flash premium yes 1M
gpt-5.3-codex [Premium] GPT-5.3 Codex premium yes 400K
gpt-5.4 [Premium] GPT-5.4 premium yes 400K
gpt-5.4-mini [Premium] GPT-5.4 Mini premium yes 400K
gpt-5.5 [Premium] GPT-5.5 premium yes 400K
poolside/laguna-s-2.1-free [Free] Laguna S 2.1 open-source yes 256K
inclusionai/ling-3.0-flash-free [Free] Ling 3.0 Flash open-source yes 256K
minimax/minimax-m2.7-free [Free] MiniMax M2.7 open-source no 197K
MiniMaxAI/MiniMax-M3-Free [Free] MiniMax M3 open-source yes 1M
minimax/minimax-m3-free [Free] MiniMax M3 open-source yes 1M
stealth/ox-alpha [Free] Ox Alpha open-source yes 1M
tencent/Hy3 [Free] Tencent Hy3 (Free) open-source yes 262K
deepseek/deepseek-v4-flash DeepSeek V4 Flash (latest) open-source yes 1M
deepseek/deepseek-v4-flash-vision-exp DeepSeek V4 Flash Vision (exp) open-source yes 1M
deepseek/deepseek-v4-pro DeepSeek V4 Pro (latest) open-source yes 1M
sakana/fugu-ultra Fugu Ultra open-source yes 1M
zai-org/GLM-5 GLM-5 open-source no 200K
zai-org/GLM-5.1 GLM-5.1 open-source no 200K
zai-org/GLM-5.2 GLM-5.2 open-source yes 1M
zai-org/GLM-5.2-Fast GLM-5.2 Fast open-source no 1M
zai-org/GLM-5.3 GLM-5.3 open-source yes 1M
gpt-5.6-luna GPT-5.6 Luna open-source yes 1M
gpt-5.6-sol GPT-5.6 Sol open-source yes 1M
gpt-5.6-terra GPT-5.6 Terra open-source yes 1M
xai/grok-4.5 Grok 4.5 open-source yes 500K
xai/grok-4.6 Grok 4.6 open-source yes 500K
thinkingmachines/inkling Inkling open-source yes 256K
thinkingmachines/inkling-small Inkling Small open-source yes 1M
moonshotai/Kimi-K2.5 Kimi K2.5 open-source no 256K
moonshotai/Kimi-K2.6 Kimi K2.6 open-source no 256K
moonshotai/Kimi-K2.7-Code Kimi K2.7 Code open-source yes 256K
moonshotai/Kimi-K2.7-Code-Highspeed Kimi K2.7 Code HighSpeed open-source yes 262K
moonshotai/Kimi-K3 Kimi K3 open-source yes 1M
xiaomi/mimo-v2.5 MiMo V2.5 open-source no 1M
xiaomi/mimo-v2.5-pro MiMo V2.5 Pro open-source no 1M
MiniMaxAI/MiniMax-M2.5 MiniMax M2.5 open-source no 200K
MiniMaxAI/MiniMax-M2.7 MiniMax M2.7 open-source no 1M
MiniMaxAI/MiniMax-M3 MiniMax M3 open-source yes 1M
meta/muse-spark-1.1 Muse Spark 1.1 open-source yes 1M
meta/muse-spark-1.2 Muse Spark 1.2 open-source yes 1M
meta/muse-spark-1.2-contributor Muse Spark 1.2 Contributor open-source yes 1M
nvidia/nemotron-3-ultra-550b-a55b Nemotron 3 Ultra open-source yes 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
Qwen/Qwen3.7-Flash Qwen 3.7 Flash open-source yes 1M
Qwen/Qwen3.7-Max Qwen 3.7 Max open-source yes 1M
Qwen/Qwen3.7-Plus Qwen 3.7 Plus open-source yes 1M
Qwen/Qwen3.8-27B Qwen 3.8 27B open-source yes 262K
Qwen/Qwen3.8-Max Qwen 3.8 Max 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
tencent/hy3-paid Tencent Hy3 open-source yes 262K

Full model list is maintained in models.json. See Sync Models to refresh it from the latest Command Code CLI release.

Development

git clone https://github.com/yoni13/opencode-commandcode-provider.git
cd commandcode-go-opencode-provider
bun install

For local testing, create opencode.local.json (gitignored) with file:// paths:

{
  "plugin": ["file:///path/to/commandcode-go-opencode-provider/server"],
  "provider": {
    "commandcode": {
      "npm": "file:///path/to/commandcode-go-opencode-provider",
      "name": "Command Code (local)",
      "env": ["COMMANDCODE_API_KEY"]
    }
  }
}

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

Sync Models

Refresh the model catalog whenever Command Code publishes new models:

bun install
bun run sync
bun run generate-readme
bun run typecheck

bun run sync downloads the latest command-code package from npm, extracts its model catalog and CLI pricing table, fetches public pricing from https://commandcode.ai/models, and writes models.json. Models without published pricing are still included with a $0/$0 placeholder so the catalog does not silently drop newly available models.

bun run generate-readme rebuilds the Available Models table from models.json.

To also write the refreshed model map into your local global config, run:

bun run sync:global

License

MIT