Google Veo 3

Fast Inference

REST API

Model Information

Response Time:~90 sec

Status:Active

Version:

0.0.1

Updated:3 days ago

veo-3

Live Demo

Average runtime: ~90 seconds

Input

Configure model parameters

Prompt

Text prompt for video generation

The camera hangs back and ascends to a high angle. As a sports car speeds forwards with its lights on entering the frame. The camera finishes at a rear tracking shot.

Output

View generated results

Result

Preview, share or download your results with a single click.

Each execution costs $0.5 With $1 you can run this model about 2 times.

API Reference

View Full Documentation

Prerequisites

Create an API Key from the Eachlabs Console
Install the required dependencies for your chosen language (e.g., requests for Python)

API Integration Steps

1. Create a Prediction

Send a POST request to create a new prediction. This will return a prediction ID that you'll use to check the result. The request should include your model inputs and API key.

import requests
import time

API_KEY = "YOUR_API_KEY"  # Replace with your API key
HEADERS = {
    "X-API-Key": API_KEY,
    "Content-Type": "application/json"
}

def create_prediction():
    response = requests.post(
        "https://api.eachlabs.ai/v1/prediction/",
        headers=HEADERS,
        json={
            "model": "veo-3",
            "version": "0.0.1",
            "input": {
  "generate_audio": false,
  "negative_prompt": "your negative prompt here",
  "aspect_ratio": "16:9",
  "seed": 0,
  "prompt": "your prompt here"
},
            "webhook_url": ""
        }
    )
    prediction = response.json()
    
    if prediction["status"] != "success":
        raise Exception(f"Prediction failed: {prediction}")
    
    return prediction["predictionID"]

2. Get Prediction Result

Poll the prediction endpoint with the prediction ID until the result is ready. The API uses long-polling, so you'll need to repeatedly check until you receive a success status.

def get_prediction(prediction_id):
    while True:
        result = requests.get(
            f"https://api.eachlabs.ai/v1/prediction/{prediction_id}",
            headers=HEADERS
        ).json()
        
        if result["status"] == "success":
            return result
        elif result["status"] == "error":
            raise Exception(f"Prediction failed: {result}")
        
        time.sleep(1)  # Wait before polling again

3. Complete Example

Here's a complete example that puts it all together, including error handling and result processing. This shows how to create a prediction and wait for the result in a production environment.

try:
    # Create prediction
    prediction_id = create_prediction()
    print(f"Prediction created: {prediction_id}")
    
    # Get result
    result = get_prediction(prediction_id)
    print(f"Output URL: {result['output']}")
    print(f"Processing time: {result['metrics']['predict_time']}s")
except Exception as e:
    print(f"Error: {e}")

Additional Information

The API uses a two-step process: create prediction and poll for results
Response time: ~90 seconds
Rate limit: 60 requests/minute
Concurrent requests: 10 maximum
Use long-polling to check prediction status until completion

Overview

Google Veo 3 is designed for high-quality text-to-video generation. It takes textual prompts and optionally a seed value to produce short videos with cinematic coherence. Google Veo 3 focuses on realism, detail, and smooth camera motion. It supports a wide range of video types such as aerial shots, slow motion, and first-person perspective videos, offering creative control through natural language descriptions.

Technical Specifications

Frame Rate: Up to 30 fps.

Resolution: Supports generation up to 1080p in preview modes.

Motion Awareness: Integrated camera and object motion tracking.

Semantic Understanding: Natural language input with multi-modal scene grounding.

Consistency: Improved temporal and spatial coherence across frames.

Key Considerations

Video duration is fixed to short clips and cannot be extended beyond a few seconds per run.

Input text is the sole control mechanism; no image, audio, or video input is supported.

Outputs may occasionally contain unnatural object deformations or flickering.

Explicit, graphic, or flagged terms may cause failure or result in blank output.

Abstract prompts may lead to hallucinated or visually ambiguous results.

Real names, brands, or sensitive entities should be avoided in prompts.

Tips & Tricks

Prompt (String):

Use clear and concrete descriptions:
✅ “A white horse galloping through a foggy forest at sunrise”
❌ “Freedom in nature”
For dynamic motion:
"A drone shot of a bustling city at night with moving cars and bright lights"
To describe visual style:
"In the style of a cinematic sci-fi film, with dramatic lighting"
Avoid flagged or unsafe content such as:
- Violence: "explosions", "war scenes"
- NSFW: "nudity", "suggestive themes"
- Identifiable individuals or real persons

Seed (Integer):

Controls variation for repeatable results.
Use values like 12345, 42, etc. for consistent video output across different generations.
Changing the seed results in slightly different visuals with the same prompt.
Leaving seed empty allows random variation.

Capabilities

Generate short cinematic-style videos from natural language.

Supports descriptions of motion, objects, scenery, and atmosphere.

Capable of handling various themes like nature, futuristic, urban, fantasy, and more.

Supports camera controls like zoom, pan, dolly, and aerial views through language.

What can I use for?

Creating short visual concepts for storytelling or ideation

Generating teaser videos for creative projects

Producing aesthetic motion clips based on detailed prompts

Visualizing scene ideas for design, animation, or narrative planning

Things to be aware of

Combine camera directions with settings:
“A slow pan across a desert at golden hour”

Mix motion and mood:
“A handheld shot following a child running through a sunflower field in slow motion”

Experiment with time of day and lighting:
“A mountain village at dusk, with lights flickering on and smoke rising from chimneys”

Add genre-based visual tones:
“Cyberpunk city with neon signs and rainy streets, drone footage”

Limitations

Realism may degrade with overly abstract prompts

May generate flickering or frame inconsistencies

No interactive editing or feedback loop — one-shot generation

Prompts involving copyrighted characters or brands may fail

Output Type: MP4

Related AI Models

You can seamlessly integrate advanced AI capabilities into your applications without the hassle of managing complex infrastructure.

Kling v1.6 Standard Text to Video

Kling v1.6 Standard Text to Video generates stable videos from text inputs, focusing on simplicity and reliability.

Seedance v1 Lite Text to Video

Seedance 1.0 Lite

Luma Ray2 720p

Generate 5s and 9s 720p videos

Minimax Hailuo T2V-01

Text transforms into clear, reliable videos using Minimax Hailuo T2V-01, built for stable and consistent video generation.