Skip to main content

OpenClaw

OpenClaw model refs use provider/model. Configure Voxvey as a custom OpenAI-compatible provider, then select a model from that provider.

1. Set the token​

export VOXVEY_TOKEN="<access-token>"

2. Add the provider​

Add this provider to your OpenClaw config.

{
  models: {
    providers: {
      voxvey: {
        name: "Voxvey",
        baseUrl: "https://api.voxvey.com/v1",
        api: "openai-completions",
        apiKey: "__env__:VOXVEY_TOKEN",
        models: [
          {
            id: "deepseek/deepseek-v4-flash",
            name: "DeepSeek V4 Flash",
            reasoning: true,
            input: ["text"],
            contextWindow: 1000000,
            maxTokens: 384000,
            compat: {
              supportsUsageInStreaming: true,
              supportsReasoningEffort: true,
              maxTokensField: "max_tokens"
            }
          }
        ]
      }
    }
  }
}

3. Select the model​

Because the Voxvey model ID contains a provider prefix, the full OpenClaw model ref includes both prefixes:

{
  agents: {
    defaults: {
      model: {
        primary: "voxvey/deepseek/deepseek-v4-flash"
      },
      models: {
        "voxvey/deepseek/deepseek-v4-flash": {
          alias: "Voxvey DeepSeek V4 Flash"
        }
      }
    }
  }
}

4. Verify​

openclaw models list --provider voxvey
openclaw models set voxvey/deepseek/deepseek-v4-flash

Start a new OpenClaw session after changing the default model. If an existing session is pinned to another model, switch it with /model voxvey/deepseek/deepseek-v4-flash.

Responses alternative​

For a Responses-compatible OpenClaw provider, use api: "openai-responses" and register openai/gpt-5.5 instead.

{
  models: {
    providers: {
      voxvey_responses: {
        name: "Voxvey Responses",
        baseUrl: "https://api.voxvey.com/v1",
        api: "openai-responses",
        apiKey: "__env__:VOXVEY_TOKEN",
        models: [
          {
            id: "openai/gpt-5.5",
            name: "GPT-5.5 via Voxvey",
            reasoning: true,
            input: ["text"],
            contextWindow: 400000,
            maxTokens: 64000
          }
        ]
      }
    }
  }
}

Messages API alternative​

For OpenClaw clients that support Messages API transport, use api: "anthropic-messages" and set the base URL to the host root. OpenClaw will call /v1/messages.

{
  models: {
    providers: {
      voxvey_messages: {
        name: "Voxvey Messages",
        baseUrl: "https://api.voxvey.com",
        api: "anthropic-messages",
        apiKey: "__env__:VOXVEY_TOKEN",
        models: [
          {
            id: "<messages-model-id>",
            name: "Voxvey Messages API model",
            reasoning: true,
            input: ["text"],
            contextWindow: 200000,
            maxTokens: 64000
          }
        ]
      }
    }
  },
  agents: {
    defaults: {
      model: {
        primary: "voxvey_messages/<messages-model-id>"
      }
    }
  }
}

Use the model ID returned by Voxvey's /v1/models for the Messages API.