Noryva API documentation

01

How the API works

Every request goes to Noryva with your API key. Pick the base URL that matches your client:

Models fall into three groups, each with one endpoint:

The text, image and video sections end with a card for each model series: the models in it, how they differ and the parameters they take. Prices are on the Models page in the console.

02

Quickstart

1. Sign in to the console and add Credits in Wallet.
2. Create a key on the API keys page.
3. Send your first request:

export NORYVA_API_KEY="your-api-key"

curl https://noryva.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.8-flash",
    "messages": [{"role": "user", "content": "Say hello in one sentence."}]
  }'

# pip install openai
import os
from openai import OpenAI

client = OpenAI(
    base_url="https://noryva.ai/api/v1",
    api_key=os.environ["NORYVA_API_KEY"],
)

response = client.chat.completions.create(
    model="qwen3.8-flash",
    messages=[{"role": "user", "content": "Say hello in one sentence."}],
)
print(response.choices[0].message.content)

// npm install openai
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://noryva.ai/api/v1",
  apiKey: process.env.NORYVA_API_KEY,
});

const response = await client.chat.completions.create({
  model: "qwen3.8-flash",
  messages: [{ role: "user", content: "Say hello in one sentence." }],
});
console.log(response.choices[0].message.content);

A successful response contains choices[0].message.content and a usage object with the tokens that were billed.

03

Authentication

Send the key in the Authorization header: Authorization: Bearer YOUR_KEY. Anthropic-format requests can use the x-api-key header instead.

• When you create a key you can limit it to certain models, IP addresses and a spending cap, and give it an expiry date.
• Keep keys on your server. Do not put them in browser code, mobile apps or public repositories.
• If a key may have leaked, delete it in the console and create a new one. Usage made with a leaked key is billed to your account.

04

Text models

Text models use the OpenAI Chat Completions format: model, messages and the usual options such as temperature, max_completion_tokens, tools and response_format. Support for structured output and tool calling varies by model.

Streaming

Set "stream": true to receive the answer as server-sent events, ending with data: [DONE]. Add "stream_options": {"include_usage": true} to get token usage in the last chunk.

curl https://noryva.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.8-flash",
    "messages": [{"role": "user", "content": "Write a haiku about cedar trees."}],
    "stream": true,
    "stream_options": {"include_usage": true}
  }'

Reasoning

Reasoning models return their thinking in a separate reasoning_content field next to content. How to switch thinking on or off, or limit it, depends on the model (for example enable_thinking or reasoning_effort); see the model's card below.

Anthropic format

Clients that speak the Anthropic Messages API, such as Claude Code, can use https://noryva.ai/api as their base URL with any text model. Requests are translated to each provider's format; options that only Anthropic models offer may be ignored.

Long requests

A request can stay open for up to 630 seconds. For long answers, use streaming or raise your client timeout.

curl https://noryva.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4-pro-0813",
    "messages": [
      {
        "role": "user",
        "content": "List three benefits of unit tests."
      }
    ],
    "enable_thinking": false
  }'

curl https://noryva.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3.8-max",
    "messages": [
      {
        "role": "user",
        "content": "List three benefits of unit tests."
      }
    ],
    "reasoning_effort": "medium"
  }'

05

Image models

Image models use POST /v1/images/generations with a JSON body. The request returns when the images are ready, which can take a few minutes, so allow at least 600 seconds on the client.

curl https://noryva.ai/api/v1/images/generations \
  --max-time 600 \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen-image-3.0",
    "prompt": "A red paper boat on a calm lake at sunrise, soft light",
    "size": "1024x1024",
    "n": 1
  }'

• The response lists the images in data[].url. These links are temporary, from 24 hours to about 7 days depending on the model, so download the files you want to keep.
• Accepted sizes, the number of images per request and how to send input images for editing differ by model. Check the model's card below before you send them.

curl https://noryva.ai/api/v1/images/generations \
  --max-time 600 \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen-image-3.0",
    "prompt": "A red paper boat on a calm lake at sunrise, soft light",
    "size": "1024x1024",
    "n": 1
  }'

06

Video models

Video generation runs as a task: create it, check it until it finishes, then download the file. It usually takes one to a few minutes.

# 1. Create the task
curl https://noryva.ai/api/v1/videos \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kling-video-v3",
    "prompt": "A red paper boat drifting on a calm lake at sunrise",
    "seconds": 5,
    "size": "1280x720"
  }'
# => {"id": "video_...", "status": "queued", ...}

# 2. Check the task every 10 to 15 seconds
curl https://noryva.ai/api/v1/videos/VIDEO_ID \
  -H "Authorization: Bearer $NORYVA_API_KEY"

# 3. When "status" is "completed", download the MP4
curl https://noryva.ai/api/v1/videos/VIDEO_ID/content \
  -H "Authorization: Bearer $NORYVA_API_KEY" -o video.mp4

• status moves from queued to in_progress and ends as completed or failed. A failed task includes an error with the reason.
• model, prompt and seconds work across video models. Allowed durations, how to set the resolution and aspect ratio, first frames, reference media and audio differ by model; see the model's card below.
• Download the file soon after the task completes. Links expire, after 24 hours for some models.
• If a create request times out on your side, do not resubmit straight away: the task may already be running and billed. Look for it in the console's task logs first.

curl https://noryva.ai/api/v1/videos \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "minimax-hailuo-h3",
    "prompt": "A red paper boat drifting on a calm lake at sunrise",
    "seconds": 5,
    "resolution": "768p",
    "aspect_ratio": "16:9"
  }'

curl https://noryva.ai/api/v1/videos \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kling-video-v3",
    "prompt": "A red paper boat drifting on a calm lake at sunrise",
    "seconds": 5,
    "resolution": "720p",
    "aspect_ratio": "16:9"
  }'

curl https://noryva.ai/api/v1/videos \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "doubao-seedance-2-0-fast-260128",
    "prompt": "A red paper boat drifting on a calm lake at sunrise",
    "seconds": 5,
    "metadata": {
      "resolution": "720p",
      "ratio": "16:9"
    }
  }'

curl https://noryva.ai/api/v1/videos \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "veo-3.1-lite",
    "prompt": "A red paper boat drifting on a calm lake at sunrise",
    "seconds": 8,
    "size": "1280x720"
  }'

curl https://noryva.ai/api/v1/videos \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "wan3.0-video",
    "prompt": "A red paper boat drifting on a calm lake at sunrise",
    "seconds": 5,
    "resolution": "720P",
    "ratio": "16:9"
  }'

curl https://noryva.ai/api/v1/videos \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tencent-video-enhance",
    "video": "https://example.com/clip.mp4",
    "seconds": 5,
    "resolution": "1080p"
  }'

curl https://noryva.ai/api/v1/videos \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tencent-video-hdr",
    "video": "https://example.com/clip.mp4",
    "seconds": 5,
    "hdr_type": "hdr10"
  }'

curl https://noryva.ai/api/v1/videos \
  -H "Authorization: Bearer $NORYVA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tencent-audio-separate",
    "video": "https://example.com/clip.mp4",
    "seconds": 30,
    "scene": "music",
    "track": "background"
  }'

07

Credits and billing

Noryva is prepaid. Usage is deducted from your Credits balance, where 100 Credits equal US$1. Top up in Wallet.

• Each model's rates are on the Models page. Some models charge more at certain hours; their page shows when.
• Image and video requests reserve Credits when they start; the final charge follows the actual result, and failed tasks are refunded.
• Every request and its cost appear in Usage logs.
• Refunds of purchases follow the Refund Policy.

08

Errors and retries

Errors use the OpenAI format:

{
  "error": {
    "message": "A readable description of the problem",
    "type": "...",
    "code": "insufficient_user_quota"
  }
}

Content checks

Model providers check prompts, input images and generated results against their content policies. An image or video task that is blocked ends as failed, with an error such as ContentModerationFailed, InvalidParameter.ViolationContent, output new_sensitive or DataInspectionFailed, and its Credits are refunded in full. These checks cannot be switched off through the API, and options such as enable_safety_checker are ignored. Reword the prompt or use different input images, then try again.

09

SDKs and tools

OpenAI SDKs and compatible tools. Set the base URL to https://noryva.ai/api/v1, use your Noryva key as the API key, and enter model IDs exactly as they appear on the Models page. This works for the official OpenAI SDKs and for tools that let you set a custom OpenAI endpoint.

Claude Code. Point it at Noryva with environment variables, then start it as usual:

export ANTHROPIC_BASE_URL="https://noryva.ai/api"
export ANTHROPIC_AUTH_TOKEN="$NORYVA_API_KEY"
export ANTHROPIC_MODEL="qwen3.8-max"
claude

Set ANTHROPIC_MODEL to any text model ID from the Models page.

10

Support

Questions about the API, billing or a failed request: write to service@appcpi.com or use the contact page. Include the time of the request and the model ID; never send your API key.