BlueDoAI（小蓝助手）
Measured metrics
- Overall score
- 7.2 / 10 ★★★⯨
- Stability band
- Very stable
- Uptime
- 100% （Measured here）
- Response latency
- 125 ms
- Model count
- 10
- Backend engine
- —
- Verification
- Verified
- Last checked
- 2026-10-19
- Outbound clicks
- 55 clicks
Uptime basis: Measured here, from 29 valid checks over the last 30 days, per-day reachability (not per request). The resolution is one day — 1% means roughly 0 unreachable days out of 30.
Published rates & minimums
The figures below are published by the service itself on public channels. We only log them, and they are not our measurements — always confirm on the service's own billing page.
Per-model, per-group pricing in RMB; live model plaza governs. Official supply groups run 0.45X-0.95X of each reference price (not a flat discount; verified 2026-10-19).
Billing model: Pay-as-you-go
A rate of 1.0 means official list price; 0.3 means you pay 30% of it. Not sure how the maths works out? Run the rate calculator →
About this service
BlueDoAI is operated by Chengdu Blue Elite Information Technology Co., Ltd. and provides compute for domestic Chinese models, a unified API and account-level usage management. Users pick from the text and video models that are formally open today; model permissions, RMB prices and billing units follow the live model plaza and the account configuration. Its companion product, Xiaolan Assistant Blue, is a Windows execution agent: the tool itself is free, while BlueDoAI-hosted models are billed by actual usage. Enterprises can ask about corporate settlement, invoicing, on-premise deployment, training and technical services.
Strengths
- Clearly identified operator with a full compliance stack: ICP filing, a value-added telecom business licence (B2-20230227) and a human-resources service licence. The catalogue publishes model IDs and billing units group by group, and a dedicated Billing and Refund page documents metering, price-change notice, invoicing and refund timelines. It also states that its blue1/blue2 endpoints return simulated responses for integration testing, not real inference.
Watch out for
- Models are concentrated in domestic series (DeepSeek, Kimi, Zhipu GLM, Qwen) plus Doubao video
- no GPT, Claude or Gemini models, and no open image, speech or embedding models were found. Our check covered API reachability only - a keyless GET /v1/models returns 401 - so output quality, billing accuracy and long-term stability remain untested. Group settlement coefficients change with model and time, so the live model plaza is the only price reference.
Our take: Best suited to teams working mainly with domestic text models that also need a documented corporate operator and invoicing. Compare model by model: within the official supply groups the published settlement coefficients differ (verified 2026-10-19: DeepSeek V4 Pro and V4 Flash 0.45X, V4.1 Flash 0.53X, Kimi K3 0.95X, GLM 5.2 0.50X, GLM 5.3 0.85X, Doubao Seedance 2.0 and 2.5 0.90X). These coefficients are set against the published reference price of each model and are not a flat discount. Video is settled on actual output tokens, with up to 100,000 tokens reserved at submission. The score reflects public compliance documents, documentation completeness and price transparency; no live model calls were made, so output quality, billing accuracy and long-term stability are unverified.
Model families supported
Show the 10 models the endpoint probe detected (first 10)
deepseek-v4-flashdeepseek-v4-prodeepseek-v4.1-flashdoubao-seedance-2.0doubao-seedance-2.5glm-5.2glm-5.3glm-5.3-flashkimi-k3qwen3.8-flash Daily monitoring log
| Date | Uptime | Avg latency | Samples | Status |
|---|---|---|---|---|
| 2026-10-27 | 100.00% | 125 ms | 1 | 200 |
| 2026-10-26 | 100.00% | 143 ms | 2 | 200 |
| 2026-10-25 | 100.00% | 132 ms | 2 | 200 |
| 2026-10-24 | 100.00% | 129 ms | 2 | 200 |
| 2026-10-23 | 100.00% | 149 ms | 2 | 200 |
| 2026-10-22 | 100.00% | 130 ms | 2 | 200 |
| 2026-10-21 | 100.00% | 139 ms | 2 | 200 |
| 2026-10-20 | 100.00% | 130 ms | 2 | 200 |
| 2026-10-19 | 100.00% | 143 ms | 2 | 200 |
| 2026-10-18 | 100.00% | 138 ms | 2 | 200 |
| 2026-10-17 | 100.00% | 160 ms | 2 | 200 |
| 2026-10-16 | 100.00% | 141 ms | 2 | 200 |
| 2026-10-15 | 100.00% | 134 ms | 2 | 200 |
| 2026-10-14 | 100.00% | 143 ms | 2 | 200 |
Data sources
站点自主提交 + 站方来函更正 + 编辑核验 2026-10-19；仅连通性探测（免 key /v1/models 返回 401），未做实际模型调用与扣费验证；依据站方公开的 model-catalog、model-plaza、api-access、legal/billing 页面
We are not affiliated with BlueDoAI（小蓝助手） and do not act as its agent, and we do not vouch for its service or the safety of your funds. Figures may lag or contain errors — if you spot one, please report it.
Similar services
稳明光语纪
An AI API relay hosted in mainland China and reachable without a VPN. It covers the Zhipu GLM, K…
Novita AI
Its own site describes it as: Access 200+ AI models with one API. Launch secure agent sandboxes…
TiMi CC
TiMi CC provides direct-in-China access to AI model APIs, without a VPN or an overseas card. Mod…
Yuegle API
A China-based AI API relay offering domestic access to the major model families. Claude, GPT, Ge…
api.gpt.ge
A China-based AI API relay offering domestic access to the major model families. Claude, GPT, Ge…
api.vveai.com
An AI API relay hosted in mainland China and reachable without a VPN. It covers the Claude, GPT,…