Use DaoXE in OpenClaw

OpenClaw's Gateway resolves every agent model from ~/.openclaw/openclaw.json — JSON5, hot-reloaded. Its custom-provider docs name DaoXE's shape exactly: api: "openai-completions" behind a /v1 base URL, the documented route for any OpenAI-compatible server (cheapest APIs guide).

Why route OpenClaw through DaoXE#

• One key behind a whole assistant. Chat, cron jobs, sub-agents and tool loops all resolve through the providers you declare — a DaoXE key covers hundreds of models from ~25 providers, so a large planner model and a cheap worker model share one balance.
• You own the metadata. Context window, input modalities and output caps are declared in your config, not guessed by the client.
• Reviewable by design. One JSON5 block states exactly what OpenClaw sends where — no hidden routing between the assistant and the endpoint.

Set it up in OpenClaw#

1. Register the provider. Under models.providers in ~/.openclaw/openclaw.json, add a daoxe entry: baseUrl: "https://daoxe.com/v1", api: "openai-completions", apiKey: "${DAOXE_API_KEY}" — the env-substitution form the docs use.
2. Register the models too. Add models: [{ id: "YOUR_EXACT_MODEL_ID" }] copied verbatim from DaoXE's GET /v1/models, and mark vision-capable models input: ["text", "image"]. A provider entry with no explicit model rows registers nothing.
3. Pick the defaults. Set agents.defaults.model.primary to "daoxe/YOUR_EXACT_MODEL_ID" — references are always provider/model. Optional aliases and per-model settings go under agents.defaults.models.
4. Save and verify. The Gateway watches the file and hot-reloads; or apply additive edits with openclaw config set models.providers.daoxe '<json>' --strict-json --merge (destructive rewrites need --replace). Confirm with openclaw models list, switch via openclaw models set daoxe/….

// ~/.openclaw/openclaw.json — add DaoXE as a custom provider
{
  env: { vars: { DAOXE_API_KEY: "sk-you…-key" } },
  agents: {
    defaults: {
      model: { primary: "daoxe/YOUR_EXACT_MODEL_ID" },
    },
  },
  models: {
    mode: "merge",
    providers: {
      daoxe: {
        baseUrl: "https://daoxe.com/v1",
        apiKey: "${DAOXE_API_KEY}",
        api: "openai-completions",
        models: [
          { id: "YOUR_EXACT_MODEL_ID", name: "DaoXE primary",
            input: ["text", "image"], contextWindow: 200000 },
        ],
      },
    },
  },
}

Verify you actually get the model#

Prove the endpoint works before blaming the client — if this fails, no setting will fix it:

export DAOXE_API_KEY="your_api_key"

# List the exact model IDs your account can call
curl --fail-with-body --show-error --silent \
  https://daoxe.com/v1/models \
  -H "Authorization: Bearer ${DAOXE_API_KEY}"

Prove the endpoint outside the assistant first, then diff a hard prompt against the official API at temperature 0:

curl --fail-with-body --show-error --silent \
  https://daoxe.com/v1/chat/completions \
  -H "Authorization: Bearer ${DAOXE_API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "YOUR_EXACT_MODEL_ID",
    "max_tokens": 64,
    "messages": [{"role": "user", "content": "Say hello in one sentence."}]
  }'

Frequently asked questions#