AI gateway preview

Token access for multiple AI models.

Buy a token, choose a model, and use AI through one simple console. 购买 Token 后,即可在控制台选择模型并使用。

Token-based access OpenAI-compatible mode Browser-only preview
Token console Model catalog Routing controls Usage logs

Product logic

Package tokens control the access path

ApiShift keeps the access model simple: payment creates a token, the token carries package limits, and requests are checked before model access.

01

Customers receive an ApiShift access token after payment.

02

Token controls package limits.

03

Console verifies token before model requests.

04

API endpoint can be used by developers.

Features

A compact control surface for AI access testing

The prototype focuses on the core product structure: token connection, model choice, routing preference, package preview, and developer documentation.

01

Token connection

Preview how an ApiShift access token is connected before requests.

02

Playground testing

Enter a prompt, choose a model, and see the planned result area.

03

Neutral model catalog

Review model options by task type without using external platform names.

04

Route selection

Switch between balanced, lower-cost, faster, and higher-success routes.

05

Billing preview

Show package cards and checkout intent without active payment.

06

API reference

Use concise Python and JavaScript examples for planned integration.

Access flow

How it works

ApiShift is designed around a clear package-to-token path that can later connect payment, token delivery, and request verification.

Step 1

Buy token

Select the token package that matches your testing scope.

Step 2

Receive access token

The planned flow delivers a token after checkout confirmation.

Step 3

Choose model or agent

Select a preview model or task-focused agent in the console.

Step 4

Use in console or API

Connect the token, choose a route, and test requests safely.

Model coverage

Preview routes for model aggregation

These categories show the intended product surface. Planned models are not presented as active upstream routes.

Preview

Chinese models

DeepSeek, Qwen, Doubao, Kimi, GLM, MiniMax routes.

Planned

Global models

OpenAI-compatible, Claude-compatible, Gemini-compatible, and Grok-compatible routes.

Planned

Image models

Image generation, image editing, and vision understanding.

Planned

Video models

Video generation and video understanding workflows.

Planned

Voice models

Speech to text, text to speech, and voice chat.

Preview

Agents

Code, writing, translation, data, support, and study templates.

Pricing

Preview token packs

Pricing cards show the planned product structure for token-based testing. Checkout is not active in this static prototype.

Test Token

For a small compatibility check.

$1

  • Token connection preview
  • Short prompt testing path
  • Basic usage structure
Checkout preview

Bot Token

For bot-style reply tests.

$19

  • Bot reply model option
  • Usage log preview
  • Timeout and route controls
Checkout preview

Growth Token

For broader route and model testing.

$49

  • Model center preview
  • Agent workflow preview
  • Usage and API setup cards
Coming soon

Tokens are not active yet. Payment and automatic token delivery will be added in the next stage.

FAQ

Preview-stage answers

What is an ApiShift token?

It is the access credential used by the console and API preview to check package balance before requests.

Are all models active now?

No. Models marked Preview or Planned are product-surface entries until upstream routing is activated.

Can I use API directly?

The preview includes POST /chat and an OpenAI-compatible endpoint shape for planned integration.

What happens when credits run out?

Requests should stop with an insufficient credits message, and the user can recharge when payment is active.

Disclaimer: ApiShift is an independent setup and integration service, not an official OpenAI service. This static prototype does not process payment, issue active tokens, or call a real model API.