100% official endpoints from Chinese providers

One API to China's flagship models
Official sources, no markup, no content logs

10+ of China's leading providers, official API prices with no markup, and smart routing that selects the optimal path in real time.

Alibaba Cloud
Tencent Cloud
Volcengine
Baidu AI Cloud
Zhipu AI
Moonshot AI
MiniMax
DeepSeek
Sign up freeRead the API docs
OpenAI / Anthropic compatible · Zero migration cost
MENUOverviewAPI keysModels & pricingOnline playgroundRequest logsRecharge center
☰⌕   Search features or docs◐　♢　FC
Account balance$1,284.60↑ 18.2%
Requests428,590↑ 12.8%
Service healthAll systems operational
99.72%
Usage trend
☰•••
Account balance$1,284.60↑ 18.2%
Requests428,590↑ 12.8%

One API key to the official endpoints of China's leading providers

AlibabaCloudAlibaba CloudTencentCloudTencent CloudVolcengineVolcengineBaiduCloudBaidu AI Cloud
Calling Chinese models from abroad

Three ways in, three costs

Connecting to each provider directly, a generic relay service, and OpenRouter each carry their own price once you reach production.

Direct connections: a high integration cost

A separate account, top-up, and reconciliation for every provider, and no alternative path when one of them fails.

Generic relay services: an opaque upstream

Unverifiable sources, no way to trace where your content goes, and little assurance of stability or long-term operation.

OpenRouter: a fee on every top-up

Models bill at list price, but deposits carry an extra platform fee of about 5.5%.

The InOneAPI approach

One entry point to the official sources

Upstream is 100% official endpoints from Chinese providers, billed at their unit prices, with zero content logging.

Several providers, automatic failover

Leading providers carry the same model together, so traffic moves on its own when one of them misbehaves.

Official sources, verifiable per request

Every request carries a trace ID, so the provider it hit, its timing, and its usage can all be checked.

No markup, no top-up fees

Billed at official unit prices, and price-first routing picks the lowest-priced provider for the same model.

Core capabilities

Zero migration, straight into production

Compatible with the protocols and tools you already use, with routing, billing, governance, and privacy in place from the start.

China's flagship models, every modality

DeepSeek, Qwen, Kimi, GLM and more under one integration, covering text, image, video, and audio.

Multi-factor scoring, live routing

Price, first-token latency, throughput, and error rate are scored together to pick the best path per request.

Official unit prices, no markup

Billed at the official unit price of the provider you hit, with tokens metered locally and every line verifiable.

Team governance and budgets

Workspaces, projects, and sub-keys with layered access, and budgets, model restrictions, and rate limits per key.

Redundancy across providers

Several providers serve the same model, so routes are down-weighted as performance drops and dropped on failure.

Zero content logging by default

Prompts and responses are processed in memory only, never written to a database, with nothing to switch on.

Connection pooling that lowers first-token latency

Price, speed, and balanced routing strategies

Native OpenAI and Anthropic protocol support

Request-level trace diagnostics

Proactive circuit breaking on abnormal traffic

Error-rate and budget alerts

RPM / TPM and concurrency limits

One-click setup for Claude Code and Codex

24/7 human support

How it works

The gateway decides on every request

Smart routing

Routing per request, against your goal

Price, first-token latency, throughput, and error rate are assessed live. The provider changes; the model never does.

Price-first

Picks the lowest-priced available provider for the same model, so no one compares by hand.

Speed-first and balanced

Routes lose weight as performance drops, and sustained failures move to a healthy provider.

Transparent billing

No premium on routing, costs you can check

Billed at the official unit price of the provider that served the request, with tokens metered locally by each provider's own algorithm.

Request-level tracing

X-Gateway-Trace-ID links routing, the upstream call, and metering together.

Line-by-line reconciliation

Every charge maps to an official price, so nothing has to be estimated.

Data and privacy

Content never lands on disk, and metadata exists only for billing

Prompts and responses are streamed and processed in memory only, and never written to a database.

Zero content logging is the standard

Only usage, cost, timing, and the trace ID are kept, for billing and troubleshooting.

Circuit breaking with live alerts

Runaway agents, retry storms, and key abuse are cut off at once and reported by email.

Zero-migration integration

Change one baseURL, reach China's flagship models

Both the OpenAI and Anthropic protocols are supported, so your existing SDKs and agent tools only need a new baseURL and API key.

Read the integration docs
quickstartCode examples
# -H "X-APP-NAME: Your App Name"
# -H "X-APP-URL: https://your-app.example.com"
curl "https://api.inoneapi.com/v1/chat/completions" \
  -H "Authorization: Bearer $INONEAPI_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{
  "model": "YOUR_MODEL_ID",
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ]
}'

China's flagship models, one API away

Pay as you go at official prices with no markup, and no content stored.

Sign up free