OpenClaw Installation & Configuration Guide

Connect OpenClaw to DMC: Install via curl or npm, save your token to ~/.openclaw/.env, configure OpenAI/Anthropic/Google providers, and access 500+ models.

Use DMC as the model provider for OpenClaw—a single API to access 500+ models.

Prerequisites

Node.js ≥ 22 — Install via nvm recommended

DMC account with API key enabled — Get one here

Installation

Get Your DMC Key

Install or Update OpenClaw

Recommended Method — Use the official installer:

Or install via npm:

Then run the setup wizard:

During initialization, if you want to manually configure DMC API, select Skip at the Model/Auth step. After initialization is complete, verify the gateway is running:

Configuration

Save Your DMC Key

Write your key to ~/.openclaw/.env so the daemon can use it without hardcoding in the config file:

Configure DMC Providers

DMC supports three API formats. Add one or more provider blocks to ~/.openclaw/openclaw.json:

Provider	API Format	Endpoint Called	Base URL
`DMC-openai`	`openai-completions`	`POST /v1/chat/completions`	`https://dmc.cc/v1`
`DMC-openai-responses`	`openai-responses`	`POST /v1/responses`	`https://dmc.cc/v1`
`DMC-claude`	`anthropic-messages`	`POST /messages`	`https://dmc.cc/v1`

{
  "agents": {
    "defaults": {
      "model": {
        "primary": "DMC-claude/your-claude-model-id"
      }
    }
  },
  "models": {
    "mode": "merge",
    "providers": {
      "DMC-openai": {
        "baseUrl": "https://dmc.cc/v1",
        "apiKey": "${DMC_KEY}",
        "api": "openai-completions",
        "models": [{ "id": "your-chat-model-id", "name": "OpenAI-compatible text model" }]
      },
      "DMC-openai-responses": {
        "baseUrl": "https://dmc.cc/v1",
        "apiKey": "${DMC_KEY}",
        "api": "openai-responses",
        "models": [{ "id": "your-responses-model-id", "name": "Responses API model" }]
      },
      "DMC-claude": {
        "baseUrl": "https://dmc.cc/v1",
        "apiKey": "${DMC_KEY}",
        "api": "anthropic-messages",
        "models": [{ "id": "your-claude-model-id", "name": "Claude family model" }]
      }

You can keep only the providers you need, or keep all four and use `provider/model` references to route between different formats.

Replace each `your-...-model-id` placeholder with the current model ID from the [DMC Models page](https://dmc.cc/models/) corresponding to the provider block.

<Note>
  For models requiring OpenAI Responses API, use `DMC-openai-responses` (`api: "openai-responses"`). For models using the traditional chat completions path, use `DMC-openai` (`api: "openai-completions"`). Both share the same `baseUrl` (`https://dmc.cc/v1`); the only difference is which downstream endpoint OpenClaw calls.
</Note>

<Note>
  This configuration does **not require** an `auth-profiles.json` file. Provider keys are resolved via the `${DMC_KEY}` environment variable from `models.providers.*.apiKey`. You only need `auth-profiles.json` when you want per-profile credential rotation between multiple keys.
</Note>
</Step>
</Steps>

## Verification

<Steps>
<Step title="Restart and Verify">
After adding `.env` or changing provider configuration, restart the gateway:

```bash theme={null}
openclaw gateway restart

Then run diagnostics:

A successful configuration should show:

openclaw models status will resolve the primary model to DMC-claude/your-claude-model-id and DMC providers should not show authentication warnings

openclaw models list --provider DMC-claude will list your configured models

openclaw doctor will report no configuration or verification errors for DMC providers