Investor Matt Shumer recommends Seedance 2.0, Hyperframes, Remotion, and Claude for AI video generation
Abel founder Daniel Francis requested the tool recommendations.
Positive users praise tools like Seedance, Magnific, and Runway as top options for AI video generation, while negative users dismiss some recommendations and criticize low visual quality.
No Digg Deeper questions have been answered for this story yet.
Most Activity
SeeDance 2 still the best for text to video and image to video. Realistic and animated styles. Midjourney still great for stylised image exploration. GPT Image 2 and Nano Banana 2/Pro are great for image editing. Something like this is a quick Midjourney/SeeDance exploration. One 15 second generation.
@growing_daniel "can you guys show me how to be lame"
@growing_daniel Currently it's generate in Seedance and edit in capcut. The more ambitious the more editing required.
@growing_daniel Seedance 2.0 for pixel video generation, Hyperframes and Remotion with Claude models for programmatic videos.
What's the current tool set for making AI generated videos. I know you guys are doing it now I want to make one
@growing_daniel Try http://fal.ai 🎥 search for Seedance 2.0 text to video
@growing_daniel 1. Jimeng (seedance) by ByteDance 2. Kling by Kuaishou 3. upstream by bilibili (Chinese phone # required$
@growing_daniel Wan2.2 is pretty good for local. Seedance 2.0 is current king but not open source. LTX2.3 is somewhat decent but I’ve had better experiences with Wan
@growing_daniel comfyui can be 1) a fucking pain but gives 2) unparalleled flexibility. any professionally generated content that has a repeatable content angle is made there, and codex can bang it out without much trouble. if you're really lazy, the templates are all you need
@growing_daniel it’s literally all in @runwayml bro
@growing_daniel Before touching any tools, define the cinematography and visual identity with Midjourney or Nano Banana. Skipping this step is how you end up with generic, average videos.
Generate a image with flux 2 or nano banana (best realism and allow alot of reference images) Then give that to a video model (kling and wan have some decently cheap ones) This lets you have more control over it than if you just tried to prompt video models. Theres video models that allow you to do first frame last frame images which are useful if you want to try to compose longer shots you can kinda stitch things together like that and have continuous shots, but if the first frame and last frame images are too different the model will do crazy shit to try to make the transition. Replicate is nice since you can mix and match models in single interface.
@growing_daniel Try this:
@growing_daniel try HeyGen!
@_claudiazhao Are you an unbiased advisor in this matter
@growing_daniel this is the way
@growing_daniel Higgsfield but if you don’t want to pay extra for the platform I would have claude code make a python script to hit the seedence 2 API and customize it to your liking
@growing_daniel Runway’s pretty good
@growing_daniel Heygen. Happy to hook you up with some free credits @HeyGen
@growing_daniel Hyperframes, genmedia (fal) and custom skills in your genre.
@growing_daniel I used a few AI text-to-video animation tools to make a few different versions/adaptations of a friend’s short story. Maybe I could make some creative promo content for ABEL if that’s what you had in mind.