Flixly API Video Guide
Step-by-step Flixly API instructions for generating videos with Seedance 2.0, Veo 3.1, and Kling 3.0. Includes credit costs, durations, and a working parameter table.
TL;DR
Use Flixly's named models inside the dashboard to generate 8-second 1080p videos for 10-18 credits each. Follow the eight-step workflow that starts with a reference image and ends with lip-sync export to avoid wasted generations.
Your latest 1080p video prompt just burned 45 credits and returned a 4-second clip with mismatched motion.
Standard text prompts on most platforms force repeated retries because they lack reference controls or consistent character handling.
Common friction points in video workflows
A single failed generation on a basic text-to-video call often wastes 8-12 credits and 90 seconds of queue time. Users report needing four to six iterations before motion aligns with the script when starting from raw text alone.
Limits of generic prompt approaches
Generic tools cap at 720p output and 5-second durations without built-in lip sync or reference image support. They also skip character consistency across shots, requiring manual re-uploads for every new scene.
Effective method with Flixly video models
Flixly surfaces named models including Seedance 2.0, Veo 3.1, Kling 3.0, Sora 2, and Wan 2.7. Each model accepts structured inputs such as 1080p resolution, 8-second duration, and 12 credits per generation.
Model selection by use case
- Seedance 2.0 handles dance sequences at 60 fps with 10-credit cost for 6 seconds.
- Veo 3.1 supports 1080p reference-to-video at 15 credits for 10 seconds.
- Kling 3.0 produces lip-sync output from 5-second audio clips using 8 credits.
Step-by-step generation process
- Navigate to the dashboard and select the text-to-video tool. Enter a 12-word prompt describing camera movement and subject action.
- Upload a single reference image sized 1024x576 to lock character appearance.
- Choose Veo 3.1 from the model dropdown and set duration to 8 seconds at 1080p.
- Enable the motion-poster toggle if the output needs static-to-video conversion.
- Review the credit estimate shown in the interface before submitting.
- Download the MP4 once the queue completes and inspect frame 24 for motion artifacts.
- If needed, route the output to the lip sync video page and attach a 6-second WAV file.
- Export the final file and log the total credits used for the session.
Comparison of video model parameters
| Model | Max Duration | Resolution | Credits | Reference Support |
|---|---|---|---|---|
| Seedance 2.0 | 6 seconds | 1080p | 10 | Image only |
| Veo 3.1 | 10 seconds | 1080p | 15 | Image + audio |
| Kling 3.0 | 8 seconds | 720p | 8 | Audio only |
| Sora 2 | 12 seconds | 1080p | 18 | Image + text |
Edge cases and current limits
Wan 2.7 rejects prompts exceeding 180 characters. Reference-to-video calls on reference to video require the source image to be under 2 MB. Outputs longer than 12 seconds trigger an additional 4-credit fee on most models.
The fastest path forward is to start a new generation directly in the Text to Video tool.
Frequently Asked Questions
How many credits does an 8-second Veo 3.1 generation cost?▾
Veo 3.1 charges 15 credits for an 8-second 1080p clip when a reference image is supplied.
Can I combine audio and image references in one call?▾
Yes. Select the lip sync page after the initial video render and attach both a reference image and a 6-second WAV file for an extra 4 credits.
What happens if my prompt exceeds 180 characters on Wan 2.7?▾
The generation is rejected before queuing and no credits are deducted.
Is there a limit on reference image file size?▾
Reference images must stay under 2 MB; larger files are automatically rejected at upload.
