Make a video from a few lines.
Fix any second of it.
Ads, reels, short films, explainers and music videos. Video, voice, music and text each sit on their own lane, and you change only the part you don't like.
Free while in beta. Try it without a key: a sample model runs the whole flow.
- 1
Describe it
Pick what you are making and write a few lines. Agents turn it into a script and check it.
- 2
Generate
Your own models make the video, voice, music and text, each on its own lane.
- 3
Fix what you don’t like
Select a few seconds or a part of the frame, say what should change, and keep the rest.
Questions
Which models can I use?
Video, image, voice, music and text models from Runware, OpenRouter, fal.ai, BytePlus, ElevenLabs and OpenAI. You add your own keys and pay the providers directly.
How does fixing a part work?
Select a range on the timeline, or paint over part of the picture, and describe the change. Takeloom picks a model that can do it (first and last frame, inpainting, reference images) and shows the cost first. Every attempt is a take you can switch back to.
What does it cost?
Takeloom is free while in beta. The providers bill your own keys, and you see an estimate before every generation.
Who owns what I make?
You do, subject to the terms of the model providers you use. Only use people, brands and music you have the rights to.