OAINew
gpt-6-sol
The AI model gateway for developers
One API.
One API.
Every major model.
Call GPT, Claude, Gemini, DeepSeek and more through the OpenAI interface you already know. One base URL, one key — switch models freely.
Direct access in China · Pay per use · Per-call usage records
quickstart.pyPython
# Keep your existing SDK, just point it to Keai
from openai import OpenAI client = OpenAI( base_url="https://api.keai.yun/v1", api_key="sk-your-key", ) response = client.chat.completions.create( model="deepseek-v4-pro", messages=[{"role": "user", "content": "Hello"}], )
OpenAI SDK · Chat CompletionsOne endpoint, switch models as needed
369+models online, synced in real time
OpenAI compatibleReuse your existing SDKs & tools
Direct in ChinaNo extra network setup required
Fully auditablePer-call usage & billing records
Model coverageThe models you know,
The models you know,
from one endpoint
Keai brings together leading models from China and abroad, covering chat, reasoning, coding and multimodal tasks. Models and pricing sync continuously; see the pricing page for current availability.
ANew
claude-opus-5-5GNew
gemini-3.8-flashDSNew
deepseek-v4.1-flashXNew
grok-4.7KPopular
kimi-k3QPopular
qwen3.8-maxZPopular
glm-5.3DBNew
doubao-seed-2-1-pro-260628OAINew
gpt-image-2.5-sunburstGPopular
gemini-3-pro-imageDBNew
doubao-seedream-5-0-pro-260628QPopular
qwen-image-3.0-proXPopular
grok-imagine-image-2.0DBNew
doubao-seedance-2-5-260628GPopular
veo_3_1KLPopular
kling-3.0-turboXPopular
grok-imagine-videoWPopular
wan3.0-videoMMPopular
MiniMax-Hailuo-2.3GPopular
gemini-3.1-pro-previewQPopular
qwen3.8-27b369+ models · synced in real time; see the pricing page for current versionsFull model list & pricing
Chat & reasoningText, tool calling, long context
Image generation & understandingText-to-image, editing, vision
Video generationText-to-video, image-to-video, editing
Multimodal modelsText, image and mixed inputs
QuickstartStart from your existing project,
Start from your existing project,
connected in minutes
If your project already uses the OpenAI SDK, just create a key and change the base URL. Your existing SDK calls keep working unchanged.
- Create a Keai account
Generate an API key in the console.
- Set the base URL
Point your requests to
https://api.keai.yun/v1. - Pick a model and call
Choose per task; pricing and available versions are on the pricing page.
Why KeaiSimpler integration,
Simpler integration,
clearer usage
Keai gives developers one unified way to access major LLM APIs, with protocols, connectivity, routing and billing made fully transparent.
OpenAI compatible
Keep the request format and SDKs you already know. Existing projects only need a new endpoint and key, minimizing switching cost.
Direct access in China
Reach the models you need from within China, without self-hosting overseas servers or maintaining extra network setups.
Multi-channel routing
Requests are routed across multiple upstream channels with automatic retries, reducing the impact of any single-channel outage.
Transparent pay-per-use
Billed per actual token usage. Rates are public, and the console provides per-call records you can export and reconcile.
Use casesFrom prototype
From prototype
to daily workflow
One compatible interface connects many models. Choose capabilities per task and track everything in a single console.
Agents & coding
Use code and reasoning models for completion, refactoring and multi-step task execution.
Content & translation
Pick the right model per task for writing, polishing and multilingual translation.
Data & documents
Use long-context models to organize reports, contracts and research, extracting structured information.
Multimodal apps
Call image, video and audio models on demand to fit different products and creative workflows.
FAQ
Before you start
What is Keai?
Keai is a unified AI model API gateway for developers. Through one OpenAI-compatible endpoint, you can call many major models.
How do I connect an existing OpenAI SDK project?
Set base_url to https://api.keai.yun/v1 and use an API key from the Keai console. Use the model IDs shown on the pricing page.
How many models does Keai support?
Currently 369+ models, synced in real time. Counts and available versions change as providers update; the pricing page is the source of truth.
How do I check costs and usage?
Keai bills per actual token usage. Model rates are listed on the pricing page, and the console provides per-call records with export.
What are the model capabilities and service boundaries?
Model capabilities are defined by each provider. Keai provides API access and channel routing; for data handling and terms, see the service boundary & data policy.
Spend your time on product, not plumbing.
Use the interface you already know. Call the models you need.