v20.2.0

Video Generation API Launch - 7 Premium Models Including Sora 2, Veo & Kling v2.1

Sreejan
SreejanAuthor

The long-awaited Video Generation API is now live! This major feature release introduces 7 premium video generation models from industry leaders including OpenAI, Google, Kuaishou, and Tencent. Generate stunning AI videos from text prompts with support for multiple quality levels, aspect ratios, and durations up to 8 seconds.

⚠️ Important: Responsible Usage Request

1

Please Use Resources Wisely: Video generation models have significant computational costs that we pay directly to upstream providers. To ensure sustainable service for all users:

  • Start with lighter models like Seedance Lite or Kling v2.1 for testing and iteration
  • Use premium models (Veo, Sora 2 Pro) only when high quality output is truly necessary
  • Optimize prompts before generating to minimize failed attempts
  • Use shorter durations when possible to reduce resource consumption

We appreciate your cooperation in using these resources responsibly so we can continue offering video generation capabilities to the community.

🎬 Video Generation API - New Endpoint

1

Major Platform Feature: A brand new video generation endpoint is now available, enabling text-to-video generation through a standardized API interface.

Endpoint Details

  • Endpoint: POST https://api.a4f.co/v1/video/generations
  • Authentication: Standard Bearer token authentication
  • Response Format: Base64 encoded video data (b64_json) for immediate availability and privacy
  • Availability: Pro and Ultra tiers only

Supported Parameters

  • model (string, required) - Model ID (e.g., provider-2/kling-v2.1)
  • prompt (string, required) - Text description of the desired video
  • quality (string, required) - Output resolution: 480p, 720p, 1080p
  • duration (integer, required) - Video length in seconds (1-8, model-dependent)
  • ratio (string, optional) - Aspect ratio: 1:1, 16:9, 9:16, 3:4, 4:3 (default: 9:16)

For complete API documentation, refer to the Video Generation API Reference.

Provider 2 - Video Generation Models

7

Premium Video Generation Suite: 7 state-of-the-art video generation models now available through Provider 2, spanning OpenAI Sora, Google Veo, Kuaishou Kling, and Tencent Seedance families.

OpenAI Sora Models (2 models)

  • provider-2/sora-2 - Pro, Ultra

    • OpenAI's flagship text-to-video model with excellent consistency and physics simulation
    • Max Duration: 4 seconds
    • Supported Qualities: 720p, 1080p
  • provider-2/sora-2-pro - Pro, Ultra

    • Enhanced detail version of Sora 2 with superior visual fidelity
    • Max Duration: 5 seconds
    • Supported Qualities: 720p, 1080p

Google Veo Models (2 models)

  • provider-2/veo-fast - Pro, Ultra

    • Google's fast video generation model optimized for speed
    • Max Duration: 8 seconds
    • Supported Qualities: 720p, 1080p
  • provider-2/veo - Pro, Ultra

    • Google's premium video generation model for maximum quality
    • Max Duration: 8 seconds
    • Supported Qualities: 720p, 1080p

Kuaishou Kling Model (1 model)

  • provider-2/kling-v2.1 - Pro, Ultra
    • High-quality general purpose video generation from Kuaishou
    • Max Duration: 5 seconds
    • Supported Qualities: 720p, 1080p
    • Supported Ratios: 16:9, 1:1, 9:16

Tencent Seedance Models (2 models)

  • provider-2/seedance-lite - Pro, Ultra

    • Tencent's lightweight video generation model for efficient workflows
    • Max Duration: 5 seconds
    • Supported Quality: 1080p
  • provider-2/seedance-pro - Pro, Ultra

    • Tencent's premium Seedance model optimized for dynamic motion
    • Max Duration: 5 seconds
    • Supported Quality: 1080p

Quick Reference Guide

3

Choose the Right Model: Select based on your priorities.

🏃 Longest Duration (up to 8 seconds)

💡 Recommended for Daily Use

⭐ Highest Quality Output (Use Sparingly)

Platform Impact

4
  • New Capability: Video generation joins chat, image, audio, and embedding as a core platform capability
  • Premium Access: All 7 models available exclusively on Pro and Ultra tiers
  • Multi-Provider Coverage: Video generation from 4 leading AI providers (OpenAI, Google, Kuaishou, Tencent)
  • Flexible Options: Range of models from lightweight to premium for diverse use cases

Important Notes

5
  • Tier Availability: Video generation is exclusively available on Pro and Ultra tiers. Free and Basic tier users do not have access to these models.
  • Duration Limits: Each model has specific maximum duration limits. Requesting a duration exceeding the model's limit will result in a 400 Bad Request error.
  • Quality Support: Not all models support all quality levels. Verify model capabilities before making requests.
  • Response Format: Videos are returned as Base64 encoded strings for immediate availability. The API does not store videos or return hosted URLs.
  • Documentation: Full API documentation, code examples, and integration guides available at Video Generation API Reference.