AI video editing tools Control the clip before and after it renders

Felo gives you generation-time controls instead of a timeline: a long prompt, first-frame and first-and-last-frame starts, reference images and videos, six aspect ratios, duration, resolution and sound. Revise any of them by asking in the same thread.

16:9768P6s
See the controls in action
Prompt
7,000 chars
References
9 images / 3 videos
Ratios
6
Resolution
480P–2K
Folded Wind3:4 · 8 s · 768P · first and last frame
  • Prompt
  • Input frames
  • Reference images
  • Reference videos
  • Ratio and duration
  • Resolution
  • Sound

The controls behind a Felo clip

There is no layer stack or keyframe graph. The controls are inputs to the generation: what the model starts from, what it references, how long it runs, what shape and size the frame is, and what it sounds like. Each one can be changed and re-rendered.

  • Prompt: describe the shot, camera, light, action and sound.

  • Frames: start from text, a first frame, or a first and last frame.

  • References: up to nine images and three videos to guide the look.

  • Format: six ratios, 4–30 second clips and output up to 2K.

  • Sound: ambience, effects and music directed in the prompt.

What each tool controls

These are the levers Felo exposes. Use them together; a change in one often needs a change in another.

Prompt
Up to 7,000 characters covering subject, action, camera, light, style and sound.
Input frames
Text-to-video, first-frame image-to-video, or first-and-last-frame to lock the opening and closing composition.
Reference images
Up to nine images, 30 MB each, to guide a character, product or look.
Reference videos
Up to three clips of 2–15 seconds each, 15 seconds total, as motion or style references.
Ratio and duration
21:9, 16:9, 4:3, 1:1, 3:4 or 9:16; 4–30 seconds at 480P/768P and 4–15 seconds at 1080P/2K.
Resolution
480P or 768P for drafts; 1080P or 2K for final clips.
Sound
Direct ambience, effects and music in the prompt; the clip renders with an audio track.
Glass Bloom3:4 · 8 s · 2K · first frame

Controls in practice

Use the same prompt fields to direct a new shot, match a look or fix a take.

  1. 01

    Direct a new shot

    Set subject, action, camera, light and sound before the first render.

  2. 02

    Match a reference

    Describe the palette and texture of an image or clip you already have.

  3. 03

    Fix one element

    Change the light or camera while leaving the subject and action in place.

Video models in Felo

Felo routes each shot to a video model. Available models and their limits are shown in video settings.

  • MiniMax H3

    Available now

    Text, image and first/last-frame video

    4–15 s · 2K

  • Seedance 2.5

    Coming soon

    Fast, motion-heavy short clips

    — · —

  • Wan 3.0

    Coming soon

    Flexible generation modes

    — · —

  • Gemini Omni Flash

    Coming soon

    Fast multimodal generation

    — · —

  • Kling 3.0

    Coming soon

    Cinematic motion and character consistency

    — · —

MiniMax H3 supports 21:9, 16:9, 4:3, 1:1, 3:4 and 9:16. Limits are shown in video settings and can differ by model; upcoming capabilities will be listed at launch.

How the editing tools work together

Choose the input, set the format, then direct the result.

  1. 1

    Choose the input and references

    Start from text, a first frame, or a first and last frame. Add up to nine images or three short videos as references.

  2. 2

    Set the format

    Pick a ratio, duration and resolution. Higher resolutions support shorter clips, so choose both together.

  3. 3

    Direct and re-render

    Revise the prompt until the clip is right. Each pass generates a new version with the same thread context.

Control examples

Each clip uses a different combination of tools — text, first frame, first and last frame, a ratio, a duration and a sound direction.

The Ocean Above the Crossing

First-and-Last-Frame Image-to-Video (FL2VA)
9:16 · 8 s · 2K

Glass Bloom

First-Frame Image-to-Video (I2VA)
3:4 · 8 s · 2K

Mirror Skin

First-Frame Image-to-Video
3:4 · 6 s · 768P

Citrus Orbit

First-Frame Image-to-Video
1:1 · 6 s · 768P

The Pancake Heist

Text-to-Video
9:16 · 6 s · 2K
Open the tools in FeloSee the full gallery

Is Felo the right fit?

Good fit

  • Prompt-level direction
  • Reproducible settings
  • Same-thread revisions

Not a fit

  • Frame-level keyframes
  • Timeline transitions
  • Captions or voiceover

Key takeaways

  • Felo's editing tools are generation controls, not a timeline.
  • Prompt up to 7,000 characters, with sound direction included.
  • Start from text, a first frame, or a first and last frame.
  • Up to nine reference images and three reference videos can guide a clip.
  • Six ratios, four resolution tiers and clips up to 30 seconds.

Frequently asked questions

What AI video editing tools does Felo include?
A prompt of up to 7,000 characters, first-frame and first-and-last-frame inputs, up to nine reference images, up to three reference videos, six aspect ratios, duration and resolution controls, and sound direction.
Does Felo have a timeline editor?
No. Felo controls generation and revision by prompt. Trimming, splitting, captions, voiceover and multi-clip assembly happen in a separate editor.
Can I use reference images?
Yes. You can add up to nine images at 30 MB each to guide a subject, product or visual style.
Can I use reference videos?
Yes. You can add up to three videos of 2–15 seconds each, up to 15 seconds in total. They guide motion or style; they are not cut into the output.
What is first-and-last-frame generation?
You provide an opening image and a closing image, and the model generates the motion between them.
How long can a prompt be?
Up to 7,000 characters.
What resolutions and durations are available?
480P and 768P support 4–30 seconds; 1080P and 2K support 4–15 seconds.
Can I add captions or a voiceover?
No. Add them in a video editor after downloading the clip.
How much do the video tools cost?
Video generations use Felo points. The pricing page lists the points included with each plan.
How do I control camera movement in the prompt?
Name one move and its speed: static, push in, track, arc, crane or pull out. Say what the move reveals and where the shot ends.
What is the 7,000-character prompt for?
The prompt field accepts up to 7,000 characters, enough for a shot list with subject, action, camera, light, style and sound. A longer prompt is not automatically better; keep each element specific.

Explore more Felo video tools

Put the controls to work

Start with text or a frame, then direct the shot with the prompt.

Open Felo video tools
Browse demo videos