Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.mirobody.ai/llms.txt

Use this file to discover all available pages before exploring further.

Volcengine Ark is ByteDance’s foundation-model service. It’s primarily used for the Doubao family, plus a curated set of third-party models (DeepSeek, Kimi, Qwen).

1. Register and get a key

1

Enable Ark

Sign in to Volcengine console and enable Ark (first-time real-name verification required for mainland China users).
2

Subscribe to models

In the Model Plaza, click Enable + Try Now on each model you need. Most Doubao models have a free trial; some are token-metered.
3

Create an API Key

On the API Key page, create a key (shown once).
4

(Optional) Create an Endpoint

Some models are called via “online inference Endpoints” — create an endpoint in the Endpoint list to get an ep-xxxxxxxxxxxx identifier. Mirobody calls via model id directly, so this step is not required for the standard flow.

2. Endpoint info

FieldValue
OpenAI-compatible base URLhttps://ark.cn-beijing.volces.com/api/v3
AuthAuthorization: Bearer <ARK_API_KEY>
RegionBeijing (ark+cn-beijing, primary)

3. Available models

Ark headlines the Doubao family, with curated third-party models. The Mirobody /api/chat provider field routes here. See the Ark model list for the live catalog.
Mirobody providerUnderlying model idNotes
豆包doubao-seed-2-0-lite-260428Mirobody’s current default, mature for Chinese workloads
doubao-seed-1-6Doubao Seed multimodal series
doubao-1-5-pro-32k / doubao-1-5-pro-256kDoubao 1.5 Pro, 32k / 256k context
doubao-1-5-vision-pro-32kVision understanding
doubao-1-5-lite-32kLightweight, low cost

Third-party (subscribe via Ark first)

  • deepseek-r1 / deepseek-v3.1 — reasoning / general chat
  • moonshot-v1-128k, moonshot-kimi-k2 — Kimi family
  • qwen2.5-72b-instruct and other Qwen models

4. Use it in Mirobody

curl -N -X POST https://mcp.thetahealth.cn/api/chat \
  -H "Authorization: Bearer $TOKEN" \
  -H "X-LLM-Provider: volcengine" \
  -H "X-LLM-Api-Key: <your Ark API key>" \
  -H "Content-Type: application/json" \
  -d '{"question":"hi","agent":"App","provider":"豆包"}'

Troubleshooting

  • Model not found → the model isn’t subscribed in your Ark account; go to Model Plaza and enable it.
  • AuthenticationError → API Key invalid or revoked; regenerate.
  • RateLimitExceeded → QPS / TPM limit on Ark; adjust quotas or request whitelisting.
  • Model IDs versioned with date suffix (e.g. 260428) — Mirobody maintains the active mapping; email Mirobody Support if a provider stops working.