OpenClaw with ARK

OpenClaw: An open-source AI assistant that runs on your own computer and works in the messaging channels you already use.

Before you start

You need an ARK API key: sign in to ARK Cloud and copy it from the dashboard, or use a key from your own ARK deployment and its base URL. The setup below gives each of OpenClaw's model roles an ARK Cloud model: arklabsai/Qwen3.5-122B-A10B as the primary model, arklabsai/Qwen3.6-35B-A3B for sub-agents and arklabsai/Qwen3.5-4B for short internal tasks. All three call tools; the 122B and the 35B read up to 256K tokens, the 4B up to 128K. Any chat model from the catalogue works.

1. Install

curl -fsSL https://openclaw.ai/install.sh | bash

2. Connect it to ARK

OpenClaw reads ~/.openclaw/openclaw.json (JSON5) and applies edits automatically. Add ARK as a provider with three models and give each role its model:

{
  agents: {
    defaults: {
      model: { primary: "ark/arklabsai/Qwen3.5-122B-A10B" },
      subagents: { model: "ark/arklabsai/Qwen3.6-35B-A3B" },
      utilityModel: "ark/arklabsai/Qwen3.5-4B",
    },
  },
  models: {
    mode: "merge",
    providers: {
      ark: {
        baseUrl: "https://api.ark-labs.cloud/api/v1",
        apiKey: "${ARK_API_KEY}",
        api: "openai-completions",
        models: [
          {
            id: "arklabsai/Qwen3.5-122B-A10B",
            name: "Qwen3.5 122B A10B",
            contextWindow: 262144,
            maxTokens: 8192,
          },
          {
            id: "arklabsai/Qwen3.6-35B-A3B",
            name: "Qwen3.6 35B A3B",
            contextWindow: 262144,
            maxTokens: 8192,
          },
          {
            id: "arklabsai/Qwen3.5-4B",
            name: "Qwen3.5 4B",
            contextWindow: 131072,
            maxTokens: 8192,
          },
        ],
      },
    },
  },
}

model.primary runs your sessions, subagents.model the sub-agents they spawn, and utilityModel short internal tasks such as session titles and transcript summaries.

Put ARK_API_KEY=… in ~/.openclaw/.env. Check the result with openclaw models status. You can also run openclaw onboard and choose a custom OpenAI-compatible provider.

Good to know

• OpenClaw splits model references on the first slash, so ARK model ids need the provider prefix: ark/arklabsai/Qwen3.6-35B-A3B.
• Use api: "openai-completions"; ARK's Anthropic route expects a bearer token, so the OpenAI route is the simpler choice.
• These tools send the whole conversation on every turn. ARK's prompt cache speeds up the repeated part of each prompt; every prompt token is still billed.

Next

Read about tool calling and limits and capacity on ARK, or browse other integrations.

Configuration keys follow OpenClaw's official documentation as of November 2026. If the tool has changed since, its docs take precedence.