29 Providers · 88+ Models · Live

One API. 88+ Models. every frontier model

Access DeepSeek, Qwen, GLM, MiniMax, Doubao, 29+ providers through a single endpoint. Our Auto Router picks the best model for every task — saving up to 93% vs GPT-4.

免费注册 · 领 ¥5 体验金 View Plans API Docs

新用户注册即送 ¥5 体验金 · 首充最高加赠 ¥180

琼ICP备2026009730号-1 · 海南艾创矩阵科技有限公司

Model Library

Every model, one endpoint

88+ models across 29 providers. Text generation, video creation, image synthesis — all behind a single API key.

Pricing

充值即用 · 全模型

预存余额，按量计费，用多少扣多少。充值越多，调用越快。所有模型 OpenAI 兼容 + Auto Router。

基础

充任意额

20 并发 · 首充加赠 ¥10-¥180

全模型按量计费
88+模型任选
适合轻量试用

进阶

¥100

60 并发

全模型按量计费
调用提速
适合日常使用

专业

¥500

120 并发

全模型按量计费
高速调用
适合高频开发

企业版

定制方案

专属并发 · SLA 按合同

全模型按量计费
分子Key · 团队管理 · 发票
优先路由 · 专属客户经理

企业版按需定制限速与专属支持 · 联系销售

智能缓存

更快响应 · 更低成本

语义缓存 + Provider 缓存，相似问题秒回，为您节省调用成本。

秒级响应

缓存命中

相似秒回

语义缓存

Token节省

Provider 缓存

Competitive Edge

Your AI Assistant with Auto Router

Auto Router selects the best model for every task, remembers your preferences, and saves up to 93% vs GPT-4. Your personal AI assistant that gets smarter with every request.

🧠

6 Categories

Auto task detection

⚡

0.3ms Overhead

$0.0000006 per call

🛡️

Auto Fallback

Primary fails → backup

💰

91% Avg Saving

vs GPT-4

With AICraft Auto Router

💻 Coding→ DeepSeek V3-93%

🌐 Translation/Batch→ V4-Flash-99%

🇨🇳 Chinese/Multilingual→ Qwen3.8-Max-88%

✍️ Creative Writing→ MiniMax M2.5-91%

Without

💻 Coding→ frontier models$2.50/M

🌐 Translation→ frontier models$2.50/M

🇨🇳 Chinese→ frontier models$2.50/M

✍️ Creative→ frontier models$2.50/M

Task	Primary Model	Fallback	Budget Pick	vs GPT-4
Coding	DeepSeek V3	Qwen-Max	V4-Flash	-93%
Translation/Batch	V4-Flash	Qwen-Turbo	V4-Flash	-99%
Chinese	Qwen-Max	GLM-5	Qwen-Turbo	-88%
Reasoning	GLM-5	DeepSeek V3	MiniMax M2.5	-83%
Creative	MiniMax M2.5	Qwen-Max	V4-Flash	-91%
Customer Service	Doubao Pro	Qwen-Turbo	V4-Flash	-91%

Quick Start

3 lines to integrate

Drop-in replacement for OpenAI. Works with Python, Node, Go, Rust, curl — any HTTP client.

# Install any OpenAI SDK
pip install openai

# Connect to AICraft
from openai import OpenAI
client = OpenAI(
    api_key="sk-your-key",
    base_url="https://aicraftapi.com/v1"
)

# Use the Auto Router (recommended)
r = client.chat.completions.create(
    model="auto",
    messages=[{"role":"user","content":"Write a Python quicksort"}]
)
# → Auto-routed to DeepSeek V3
# → Saved 93% vs GPT-4

Trust & Compliance

Privacy by jurisdiction

GDPR (EU) · PIPL (China) · LGPD (Brazil) · CCPA (California)

Data Commitments

No conversation content stored

TLS 1.3 encryption end-to-end

API keys cryptographically hashed

Chinese model data stays in China

International via Hainan Free Trade Port

30-day max log retention

GDPR (EU/EEA)

AICraft as Data Processor

SCCs for cross-border transfers

Right to erasure (30 days)

DPO: dpo@aicraftapi.com

PIPL (China)

受托处理者：AICraft

法律依据：合同必需+知情同意

数据跨境：海南自贸港

联系：dpo@aicraftapi.com

LGPD · CCPA

LGPD: Operador de dados · SCCs

CCPA: We do NOT sell personal data

Right to know · Right to delete

privacy@aicraftapi.com

I agree to Privacy Policy & Terms of Service.