HOUJIAYAN API

Global AI API Gateway

One API, 11+ Chinese AI models.

OpenAI-compatible endpoints for DeepSeek, Qwen, GLM, MiniMax, and Doubao. Self-serve signup, 100,000 free tokens, transparent pricing. Built in Hohhot, powered by Inner Mongolia's renewable grid.

Create free account → Browse models
🎁 100,000 free tokens · no credit card · ready in 30 seconds
11+
Models from 5 vendors
$0.15
Starting price / 1M input
99%
Uptime target
30s
From signup to first call

Models that work for your stack

Flagship reasoning, balanced chat, vision, code, and OCR — all on one OpenAI-compatible endpoint. Switch with a single parameter change.

DeepSeek
deepseek-v4-pro
Flagship reasoning model. 256K context. Top-tier performance on complex analysis, coding, and research.
$0.60 in $1.20 out
per 1M tokens
DeepSeek
deepseek-v4-flash
Fast & cost-effective reasoning. 1M context. Ideal for chatbots, content generation, and quick Q&A at scale.
$0.15 in $0.30 out
per 1M tokens
Alibaba Cloud
qwen3.7-max
Alibaba's flagship Qwen. 256K context. Balanced performance for general-purpose tasks, translation, and summarization.
$0.84 in $1.68 out
per 1M tokens
Alibaba Cloud
qwen3.7-plus
Enhanced Qwen with advanced reasoning and code generation. Great for programming, technical writing, and analysis.
$0.84 in $1.68 out
per 1M tokens
Alibaba Cloud
qwen3.5-ocr
Specialized OCR for document text extraction and image understanding. Receipts, contracts, scanned forms.
$0.50 in $1.00 out
per 1M tokens
Alibaba Cloud
qwen3.5-livetranslate
Real-time speech translation for live conversation. Meeting transcription, voice apps, simultaneous interpretation.
$1.00 in $2.00 out
per 1M tokens
MiniMax
MiniMax-M3
Flagship multimodal model with state-of-the-art capabilities. Vision + text + extended context for advanced agents.
$2.10 in $8.40 out
per 1M tokens
MiniMax
MiniMax-M2.7
Advanced model with tool calling and extended reasoning. Agent systems, multi-step planning, function execution.
$1.20 in $4.80 out
per 1M tokens
MiniMax
MiniMax-M2.5-highspeed
High-speed variant with 1M token context. Long document specialist — books, codebases, RAG corpora.
$0.90 in $3.60 out
per 1M tokens
ByteDance
doubao-seed-code
ByteDance's code generation specialist. Optimized for completion, debugging, and technical documentation.
$0.20 in $0.40 out
per 1M tokens
More coming
GLM · Kimi · Step
We're adding Zhipu GLM, Moonshot Kimi, and Step models in the next release. Existing customers get early access.
Q3 2026

Start with 100K free tokens →

Pricing that doesn't pretend

Per-token billing, no monthly minimum, no per-seat fee. Pay only for what you use. USDT support is in private beta — email us to join.

Model Input / 1M Output / 1M vs. Intl. equivalent
deepseek-v4-flash $0.15 $0.30 vs. GPT-5.4-nano: 33%
deepseek-v4-pro $0.60 $1.20 vs. GPT-5.4: 76% / 92%
qwen3.7-max $0.84 $1.68 vs. Claude Sonnet 4.6: 72% / 89%
qwen3.7-plus $0.84 $1.68 vs. Gemini 2.5 Flash: 60% / 90%
MiniMax-M3 $2.10 $8.40 vs. Claude Sonnet 4.6: 30% / 44%
MiniMax-M2.7 $1.20 $4.80 vs. Claude Haiku 4.5: -20% / 4%
MiniMax-M2.5-highspeed $0.90 $3.60 vs. Gemini 2.5 Flash: 28% / 28%
doubao-seed-code $0.20 $0.40 vs. GPT-5.4-nano: 0% / 68%

International prices from public vendor pages. Real-time comparison: api.houjiayan.com/pricing

Get started in three steps

No credit card. No phone verification. Sign up, get an API key, and make your first call in under a minute.

1

Create your free account

Self-serve signup. Email + password is all you need.

Sign up →
2

Generate an API key

Open the console and click "Create new key." 100,000 free tokens are credited automatically.

sk-hj-xxxxxxxxxxxxxxxx
3

Make your first call

OpenAI-compatible. Drop in your existing SDK and change the base URL.

curl https://api.houjiayan.com/v1/chat/completions \
  -H "Authorization: Bearer $HOUJIAYAN_KEY" \
  -d '{"model":"deepseek-v4-flash",...}'

Why teams pick HOUJIAYAN

Five reasons developers switch from direct upstream or other aggregators.

One integration

11 models, 5 vendors, one OpenAI-compatible endpoint. Change models with a single parameter.

Smart failover

Multi-channel routing per model. If upstream is slow or errors, we auto-failover to a backup key.

Renewable-powered

Compute runs on Inner Mongolia's wind + solar grid. Lower carbon than most European data centers.

99% uptime target

Live status page at status.houjiayan.com. Public 30-day rolling uptime for every component.

Transparent billing

Per-token, per-request. View real-time usage in the console. Export CSV anytime.

Privacy by design

We do not log prompts or responses. Single first-party cookie, no analytics, no tracking pixels.

Stop juggling 5 API keys.

One account, 11 models, transparent pricing. 100,000 free tokens the moment you sign up.

Create free account →