Five minutes
to connect.

J AI speaks an OpenAI-compatible API. Video is asynchronous: submit a task, then poll. Examples use the Beta domain.

Basics

Base URL: https://api.j-ai.app/v1

Auth: add Authorization: Bearer YOUR_KEY to every request. Create keys in the console and set a limit, an expiry date, a model allow-list and IP restrictions.

All amounts are in HKD and deducted from actual usage. Failed calls are not charged; video tasks are pre-authorised, settled on completion, and refunded if they fail or time out.

Text: chat completions

curl https://api.j-ai.app/v1/chat/completions \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-flash",
    "messages": [{"role": "user", "content": "Summarise this in three sentences..."}],
    "stream": true
  }'

Python (the official OpenAI SDK works as-is):

from openai import OpenAI

client = OpenAI(
    base_url="https://api.j-ai.app/v1",
    api_key="YOUR_KEY",
)

reply = client.chat.completions.create(
    model="deepseek-flash",
    messages=[{"role": "user", "content": "Hello"}],
)
print(reply.choices[0].message.content)

Streaming, tool calls and JSON mode are supported. Models: deepseek-flash, deepseek-v4-pro, MiniMax-M3, MiniMax-M3.1-Flash.

Image: generations

curl https://api.j-ai.app/v1/images/generations \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "doubao-seedream-5-5-261028",
    "prompt": "A red apple on a wooden table, soft daylight",
    "size": "1024x1024"
  }'

The returned data[].url can be downloaded directly. Image URLs may expire; save them promptly.

Voice: text to speech

curl https://api.j-ai.app/v1/audio/speech \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "speech-2.8-hd",
    "input": "Hello, this is a J AI voice test.",
    "voice": "male-qn-qingse"
  }' --output speech.mp3

Video: submit and poll

Submit a task to get a task_id:

curl https://api.j-ai.app/v1/video/generations \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "doubao-seedance-2-5-260628",
    "prompt": "Waves hitting rocks at dusk, cinematic",
    "duration": 5,
    "size": "1280x720"
  }'

Poll every 5–10 seconds until completed or failed:

curl https://api.j-ai.app/v1/video/generations/TASK_ID \
  -H "Authorization: Bearer YOUR_KEY"

Status values: queued, in_progress, completed, failed. A completed task includes a url; a failed task refunds the pre-authorised amount. J AI does not send webhooks — poll from your client.

Models

curl https://api.j-ai.app/v1/models \
  -H "Authorization: Bearer YOUR_KEY"

Prices and units are on the pricing page.

Errors

  • 401: invalid or expired key.
  • 402: insufficient balance.
  • 429: too many requests; back off and retry.
  • 5xx: upstream or platform issue; failed requests are not charged. If it persists, email hello@j-ai.app.

Beta limits

  • Each user is capped at HK$500; invite only.
  • Video tasks run 5–10 seconds, depending on the model.
  • Never put a key in browser-side code or a public repo; revoke it in the console if it leaks.
  • Prohibited uses are listed in the acceptable use policy.