Using Alibaba Model Studio with OpenClaw

Prerequisites

• ✓Access to Alibaba Model Studio / DashScope with a valid API key that can be used as MODELSTUDIO_API_KEY, DASHSCOPE_API_KEY, or QWEN_API_KEY.
• ✓An existing OpenClaw installation with the CLI available so you can run `openclaw onboard` and `openclaw models list`.
• ✓At least one Wan video model enabled in your Alibaba Model Studio account so calls to `alibaba/wan2.6-t2v` and related models succeed.

Steps

1. 1

   Choose which Alibaba credential you want OpenClaw to use

   OpenClaw’s `alibaba` provider accepts three different auth keys and checks them in a specific order. Decide whether you want to authenticate with a Model Studio key, a DashScope key, or a Qwen key so you know which secret to provision in your environment.

   textCopy

   OpenClaw ships a bundled `alibaba` video-generation provider for Wan models on Alibaba Model Studio / DashScope.
   *   Provider: `alibaba`
   *   Preferred auth: `MODELSTUDIO_API_KEY`
   *   Also accepted: `DASHSCOPE_API_KEY`, `QWEN_API_KEY`
   *   API: DashScope / Model Studio async video generation

2. 2

   Run OpenClaw onboarding with the Qwen standard API key flow

   Use the onboarding command to register your Alibaba / DashScope credential with OpenClaw. The `qwen-standard-api-key` choice configures the shared DashScope credential that the `alibaba` provider can also use.

   bashCopy

   openclaw onboard --auth-choice qwen-standard-api-key

3. 3

   Set a default Wan video model for your agents

   Configure your agent defaults so OpenClaw knows which Alibaba Wan model to call when you request video generation. This keeps your prompts clean and ensures all agents share the same primary video model unless you override it.

   jsonCopy

   {
     agents: {
       defaults: {
         videoGenerationModel: {
           primary: "alibaba/wan2.6-t2v",
         },
       },
     },
   }

4. 4

   Verify the Alibaba provider and Wan models are available

   List models for the `alibaba` provider to confirm that authentication works and the Wan models are registered. If this command returns the Wan models, your OpenClaw instance is ready to generate video through Alibaba Model Studio.

   bashCopy

   openclaw models list --provider alibaba

5. 5

   Pick the right Wan model mode for your use case

   Alibaba Wan exposes several modes (text-to-video, image-to-video, and reference-to-video) under different model refs. Choose the specific model ref that matches how you plan to drive video generation from your agents.

   textCopy

   The bundled `alibaba` provider currently registers:
   
   | Model ref | Mode |
   | --- | --- |
   | `alibaba/wan2.6-t2v` | Text-to-video |
   | `alibaba/wan2.6-i2v` | Image-to-video |
   | `alibaba/wan2.6-r2v` | Reference-to-video |
   | `alibaba/wan2.6-r2v-flash` | Reference-to-video (fast) |
   | `alibaba/wan2.7-r2v` | Reference-to-video |

6. 6

   Respect current Wan video limits when designing prompts

   Alibaba Wan has strict limits on video duration, number of inputs, and supported controls. Keep these in mind when building tools or prompts so you don’t send requests that the provider will reject.

   textCopy

   | Parameter | Limit |
   | --- | --- |
   | Output videos | Up to **1** per request |
   | Input images | Up to **1** |
   | Input videos | Up to **4** |
   | Duration | Up to **10 seconds** |
   | Supported controls | `size`, `aspectRatio`, `resolution`, `audio`, `watermark` |
   | Reference image/video | Remote `http(s)` URLs only |
   
   Reference image/video mode currently requires **remote http(s) URLs**. Local file paths are not supported for reference inputs.

Configuration

Troubleshooting

openclaw onboard --auth-choice qwen-standard-api-key

Reference image/video mode currently requires **remote http(s) URLs**. Local file paths are not supported for reference inputs.

Frequently asked questions