Internet & SaaS
Bring model integration, authentication, billing and support into one gateway. Add models without rebuilding your business integration.
TokenHub
One API for leading AI models. Standardize requests, manage access and route around failures, so your team can focus on building AI applications.
- OpenAI
- Compatible API format
- Global
- Leading model families
- Seconds
- Failover switching
One API. Many models.
OpenAI-compatible requests
POST
/chat/completionsmodel"MODEL_ID"messages[{ "role": "user", … }]
Model routing
03TextChat & reasoning
ImageVisual creation
MultimodalAudio & more
Unified response
message.contentLeading models.
One API.
Bring global AI capabilities into your workflow. Connect text, image, voice and multimodal models through one consistent API, with less integration work as your model choices grow.
OpenAI
GPT
General purpose, reasoning, coding and tool use
Anthropic
Claude
Coding, long documents and multi-step tasks
Google
Gemini
Multimodal work across images, video and audio
Meta
Llama
Open ecosystem, customization and private deployment
xAI
Grok
Fast responses and trending topics
DeepSeek
DeepSeek
Math, coding, cost efficiency and batch tasks
Alibaba
Qwen
Chinese, broad model coverage, open models and enterprise integration
ByteDance
Doubao
Chinese expression, voice, copywriting and content creation
Moonshot AI
Kimi
Long context, reports and research synthesis
Zhipu AI
GLM
Coding, tool use, agents and enterprise applications
Baidu
ERNIE
Search-augmented answers, knowledge Q&A and industry applications
Tencent
Hunyuan
Office collaboration, communities and Tencent Cloud integration
One gateway.
Four ways to build.
From your first AI feature to production agents, reduce the work of connecting and managing multiple models.
Finance & enterprise
Control model access by project and keep usage and cost records together for teams with permission and audit requirements.
AI Agent teams
Match customer service, coding assistants and automated workflows to suitable models. Monitor consumption and use fallback routes when providers fluctuate.
Channel partners
Split account quotas across customers and partners. Give each customer independent permissions and manage usage and settlement centrally.
Illustrative flows · No real requests.
One place for model access and control
0104
OpenAI-compatible API
Keep the familiar Chat Completions request format and reduce changes to your application code.
Automatic protocol conversion
The gateway adapts provider-specific APIs, while your application keeps a consistent request and response format.
Scoped sub-token access
Set quotas and available models for employees, departments and projects to reduce shared-key exposure and misuse.
Load balancing and failover
Use intelligent retries and fallback channels when a model or route fails, reducing the impact of a single point of failure.
Many models. Less to manage.
One integration, one billing cycle and one support team. TokenHub handles provider differences and routing behind the scenes.
| Comparison criteria | TokenHubOne gateway for every model | Direct to providersManage each separately |
|---|---|---|
| Integration | Connect once Access models through one OpenAI-compatible API. | Separate integrations Adapt to each provider’s API and repeat integration work. |
| Billing | One billing cycle Track model and project costs in one place. | Multiple billing cycles Reconcile separate accounts and consolidate costs yourself. |
| Support | One support team Troubleshoot authentication, routing and API calls together. | Multiple support teams Coordinate with each provider to trace issues. |
| Reliability | Automatic failover Load balancing, intelligent retries and fallback routing. | Provider-dependent A model or route failure can affect your application. |
Get connected in four steps
Set up your account, prepare resources and connect your application through a unified API.
Create an account
Register online and configure your company information.
Add resources
Prepare a resource balance for the models your application needs.
Create an API key
Choose available models and create a dedicated key or sub-token.
Call the API
Connect your application using the OpenAI-compatible request format.
DEVELOPER EXPERIENCE
A familiar API.
A simpler integration.
Use the OpenAI-compatible request format with your project API key and an available model. Start with the language your team already uses.
Contact Us
TokenHub / Chat CompletionsStatic code examples · No requests are sent
OpenAI SDK / HTTPfrom openai import OpenAIclient = OpenAI(base_url="YOUR_BASE_URL",api_key="YOUR_API_KEY",)result = client.chat.completions.create(model="MODEL_ID",messages=[{"role": "user", "content": "Hello"}],)print(result.choices[0].message.content)
import OpenAI from "openai";const client = new OpenAI({baseURL: process.env.TOKEN_BASE_URL,apiKey: process.env.TOKEN_API_KEY,});const result = await client.chat.completions.create({model: "MODEL_ID",messages: [{ role: "user", content: "Hello" }],});console.log(result.choices[0].message.content);
import OpenAI from "openai";const client = new OpenAI({baseURL: process.env.TOKEN_BASE_URL,apiKey: process.env.TOKEN_API_KEY,});const result = await client.chat.completions.create({model: "MODEL_ID",messages: [{ role: "user", content: "Hello" }],});console.log(result.choices[0].message.content);
curl "${TOKEN_BASE_URL}/chat/completions" \-H "Authorization: Bearer ${TOKEN_API_KEY}" \-H "Content-Type: application/json" \-d '{"model": "MODEL_ID","messages": [{"role": "user", "content": "Hello"}]}'
Three settings to get started
Set your project URL
Replace the Base URL with the address supplied for your environment.
Use a project key
Configure an API key or assigned sub-token on your server.
Choose an available model
Replace MODEL_ID with a model enabled for your project.
Examples use configuration placeholders. Keep API keys on your server, outside browser code.
Questions before you get started
Answers about multi-model integration and data handling.
Why use a gateway instead of separate provider integrations?
A single integration brings model access, billing and support together. The platform handles protocol differences and routing so your team can manage fewer integration paths.
BUILT TO SCALE WITH YOU
Connect AI with your customer channels
EMAIL WORKSPACE
Create templates, send through API or SMTP, and follow email events in InnoMails.
GLOBAL COVERAGE
SMS
For OTP, Promotion, and Notification, we offer a global direct connection resource channel and leverage powerful APIs to simplify integration.
COST-EFFECTIVE
We provide an officially authorized WhatsApp messaging channel to help businesses engage with users throughout the customer journey.