API Gateway

LLM Agent API Gateway for Claude Code, Codex CLI, and Chinese LLMs

Buy access once, then test GPT, Claude Sonnet 5, Claude Fable 5, Kimi, GLM, DeepSeek, StepFun, Qwen, Tencent Hunyuan, LongCat-2.0, and gpt-image-2 through compatible endpoints.

Supports a broad set of model providers

Get API accessTop up existing key
API accesshttps://gpt-agent.cc/v1
Models

Supported AI models and image generation

Coverage for developer, workflow, multimodal, and gpt-image-2 visual use cases

gpt-image-2Image generation
ClaudeLow cost
GPTPopular
GeminiMultimodal
DeepSeekReasoning
QwenCoding
DoubaoCoding
Tencent HunyuanMultimodal
KimiLong context
Zhipu GLMReasoning
Meituan LongCatAgentic coding
MiniMaxM3
StepFunReasoning
Xiaomi MiMoOn-device
More30+

Supports a broad set of model providers

Pricing
Choose your path

Buy a new key or top up an existing one

New usersChoose a plan, pay online, and receive an API key

Start here if you do not have a key yet.

  • Automatic delivery after payment
  • The first API key is sent by email
Default path for first-time buyersGo to purchase
Existing usersEnter your key, pay online, and add token quota

Top up directly when your workloads are already running.

  • Quota is added after payment
  • Track delivery status from the order page
Add quota to an existing keyGo to top-up
API access

Choose the endpoint your client expects

Use Chat Completions-compatible, Responses-compatible, or Messages-compatible endpoints. First check whether your client asks for a full endpoint or a Base URL.

OpenAI-compatible/v1/chat/completions
For common Chat Completions clients

A general compatibility endpoint for SDKs, plugins, and third-party clients that speak the Chat Completions format.

https://gpt-agent.cc/v1/chat/completions
  • Use this when a client asks for the full endpoint
  • Start here when the client type is unclear
Responses-compatible/v1/responses
For newer response-style clients

Use this endpoint when your client or workflow expects the Responses API shape.

https://gpt-agent.cc/v1/responses
  • Good for agent-style workflows
  • Works well with compatible developer tools
Messages-compatible/v1/messages
For Messages-format clients

Use this endpoint when your client expects the Messages request shape.

https://gpt-agent.cc/v1/messages
  • Messages-compatible clients
  • Use the full endpoint when required
Base URL tip

Some clients ask for a Base URL instead of a full endpoint. Start with https://gpt-agent.cc/v1, and use the site root only when the client expects it: https://gpt-agent.cc.

Recommended: https://gpt-agent.cc/v1Fallback: https://gpt-agent.cc
Service assurance

Privacy, invoices, and support are covered

This is not just an API endpoint. Purchasing, payment, delivery, and issue handling are explained before you integrate.

Privacy protection

Request content is not logged; traffic is forwarded only. See the privacy policy for details.

Read privacy policyBusiness invoices

Company invoices are supported. Contact support after payment to request one.

Ask about invoicingPost-purchase support

Contact support for delivery issues, refunds, and account questions.

View refund policy
Multi-model gateway

Access multiple model families through one interface and choose by workload.

Covers mainstream models
FAQ

Quick answers

How do I get GPT API or Claude API access?

Open the access page, pick a package, and pay online — your GPT API and Claude API credentials are provisioned automatically. One account works across GPT, Claude, DeepSeek, Doubao, Kimi, GLM and more, and you can top up usage anytime.

Can I use Claude Code with this API?

Yes. Get Claude Code access here, point Claude Code at the Messages-compatible endpoint https://gpt-agent.cc/v1, and route requests to Kimi, Doubao, GLM, MiMo, DeepSeek, or Claude models through one account.

Does Claude Code work with Kimi, Doubao, GLM, and Xiaomi MiMo APIs?

Yes. The gateway exposes Kimi API, Doubao API, GLM API, and Xiaomi MiMo API behind the same Claude-compatible endpoint, so you can run Claude Code on lower-cost models without changing your client setup.

Which API formats are supported?

Chat Completions-compatible, Responses-compatible, and Messages-compatible endpoints are available.

Is gpt-image-2 image generation supported?

Yes. gpt-image-2 is available for image generation scenarios such as ecommerce posters, product visuals, social creatives, and marketing assets.

What Base URL should I use?

Most clients can start with https://gpt-agent.cc/v1. If a client requires a full endpoint, use the path that matches its API format.

How is pricing calculated?

Most models are billed by actual token usage. Purchased quota does not expire.

Can teams request a custom plan?

Yes. High-volume teams can contact support for custom quota and support arrangements.

More questions? See FAQ or contact support