OpenClaw Using Ace Data Cloud Tutorial

OpenClaw is a multi-channel AI gateway and scalable plugin runtime that supports the integration of third-party models and tools as plugins. @acedatacloud/openclaw-provider is the official OpenClaw plugin that connects over 50 large language models from Ace Data Cloud and Google web search into OpenClaw, which can be accessed using a single API Key.

This article explains how to install, configure, and use the Ace Data Cloud plugin in OpenClaw.

Application Process

To use this plugin, you can first go to the Ace Data Cloud Console to obtain your API Token for future use.

If you are not logged in or registered, you will be automatically redirected to the login page inviting you to register and log in. After logging in or registering, you will be automatically returned to the current page.

There is a free quota available for first-time applicants, allowing you to experience the services provided by this plugin for free.

Install OpenClaw

If OpenClaw is not yet installed, you can install it globally via npm:

npm install -g openclaw

After installation, check if the command is available:

openclaw --version

Node.js version 20 or higher is required.

Install Ace Data Cloud Plugin

OpenClaw integrates third-party models through a plugin mechanism. Install @acedatacloud/openclaw-provider:

openclaw plugins install npm:@acedatacloud/openclaw-provider

After installation, OpenClaw will automatically register a acedatacloud channel, which can support both Chat and Web Search capabilities.

Configure API Token

There are three ways to configure the API Token:

Method 1: Environment Variable (Recommended)

Write the following content into ~/.zshrc, ~/.bashrc, or ~/.bash_profile:

export ACEDATA_API_KEY="{token}"

Replace {token} with the API Token you copied from the Ace Data Cloud Console. You can also use the alias environment variable ACEDATACLOUD_API_KEY, both are equivalent.

After configuration, reopen the terminal or execute:

source ~/.zshrc

Method 2: Interactive Onboarding

Run:

openclaw onboard --auth-choice acedatacloud-api-key

OpenClaw will prompt you to enter the API Token and automatically write it into the user configuration file.

Method 3: Command Line Argument

You can also pass it directly via the --acedata-api-key <key> parameter when calling.

Use Chat Capability

Once configured, you can call the LLMs provided by Ace Data Cloud in OpenClaw. The model ID format is acedatacloud/<model-name>.

One-shot Call — Use openclaw infer model run to run a single-turn conversation:

openclaw infer model run --model acedatacloud/gpt-4.1-mini --prompt "Explain what OpenClaw is in one sentence"

openclaw infer model run --model acedatacloud/claude-opus-4-8 --prompt "Explain the plugin mechanism of OpenClaw"

Set as Default Model — Subsequent interactive/Agent runs will default to using it:

openclaw config set agents.defaults.model.primary 'acedatacloud/claude-opus-4-8'

Interactive Terminal UI — Directly enter the local conversation interface:

openclaw chat

Single Agent Task — Let the local Agent run a round (requires the current shell to have the API Token set):

openclaw agent --local --model acedatacloud/claude-opus-4-8 -m "Explain what OpenClaw is in one sentence"

Default Model List

The plugin includes a directory of 62 commonly used models, covering the following families (sorted by directory):

Family Representative Models
Claude claude-opus-4-8, claude-sonnet-4-6, claude-haiku-4-5-20251001
GPT gpt-5.4-mini, gpt-5.2-pro, gpt-4.1, gpt-4.1-mini, gpt-4o
OpenAI o-series o3, o4-mini
Gemini gemini-3.1-pro, gemini-3-pro-preview, gemini-3.1-flash-lite-preview
Grok grok-4, grok-4-1-fast
DeepSeek deepseek-v4-flash, deepseek-v3.2-exp, deepseek-r1
Kimi kimi-k2.5, kimi-k2-thinking-turbo
GLM glm-5.1, glm-5, glm-4.6

For the complete list, please refer to the plugin repository's src/chat/generated-catalog.ts, which is automatically generated by pnpm sync-catalog from PlatformBackend/cost/api/_chat_models.json.

Dynamic Model ID

Model IDs not in the built-in directory can also be passed directly — the plugin will forward any acedatacloud/<id> request to https://api.acedata.cloud/v1 in an OpenAI-compatible manner via resolveDynamicModel:

openclaw infer model run --model acedatacloud/gpt-5.5 --prompt "Test new model"

As long as the id is available in the Ace Data Cloud backend (refer to /v1/models), the call will succeed.

The plugin also registers a web search provider, using Ace Data Cloud's Google SERP interface. OpenClaw will automatically call it in the Agent workflow, and it can also be explicitly called using openclaw infer web search:

openclaw infer web search --provider acedatacloud --query "Latest AI Agent frameworks" --limit 5

The available command line parameters are --provider, --query, --limit, and --json. The underlying Google SERP interface also supports search types such as search (web, default), images, news, videos, maps, places, etc., which OpenClaw will select as needed in the Agent workflow.

Billing Explanation

The plugin itself will not incur duplicate charges — all usage is calculated on the Ace Data Cloud platform based on Credits (priced by model and request size). Chat and Web Search share the same API Token, and call records can be viewed at platform.acedata.cloud/console/usages.

To avoid duplicate estimations by the OpenClaw client, the plugin reports cost: 0 to OpenClaw.

Manual Configuration

If you wish to edit the configuration file directly, you can write the API Token, default model, and enable the plugin in ~/.openclaw/openclaw.json:

{
  env: { ACEDATA_API_KEY: "ace-..." },
  agents: {
    defaults: {
      model: { primary: "acedatacloud/claude-opus-4-8" },
    },
  },
  plugins: {
    entries: {
      acedatacloud: { enabled: true },
    },
  },
}

You can also configure a fallback chain, which automatically downgrades to other models when the primary model is unavailable:

{
  agents: {
    defaults: {
      model: {
        primary: "acedatacloud/claude-opus-4-8",
        fallbacks: ["acedatacloud/gpt-5.2-pro", "acedatacloud/gemini-3.1-pro"],
      },
    },
  },
}

Troubleshooting

Phenomenon Troubleshooting Direction
Ace Data Cloud API key not configured ACEDATA_API_KEY/ACEDATACLOUD_API_KEY not set, or onboarding not completed
HTTP 401 API Token is incorrect or disabled, please regenerate it in the console
HTTP 402 Insufficient balance, please recharge at platform.acedata.cloud/console/balance
HTTP 429 Triggered upstream rate limiting, please try again later or adjust the concurrency limit in the console
Model ID not in the built-in directory but still callable This is expected behavior: dynamic ID passthrough

More debug logs can be opened with the global parameter --log-level debug, for example openclaw --log-level debug infer model run --model acedatacloud/gpt-4o-mini --prompt "test".