Over the last few months AI Video production has advanced by leaps and bounds, reaching a useful and impressive state. Here is a quick review of a few of the leading AI video production services
Hailuo AI
- Location: https://hailuoai.video/create
- Free images: 20-40 images, daily renewing
- Free Videos: 3 per day
- Pricing: $0-10-34-94
- Features: Negative prompting, image prompting, styles, models, pipelines, additional tools
- License: Commercial use allowed
Hailuo AI offer a limited amount of free generation. Additionally you can add to your standard generations and still get the daily free tier on top which feels fairer than some other platforms. The free generations frequently have a wait of 20 minutes or so to generate. Clips are 6 seconds long you can see an example here. You can add both text prompts such as:
A highly detailed and realistic professional photograph of a confident woman in her mid-30s with curly, shoulder-length brown hair, porcelain-like complexion, and subtle freckles scattered across the bridge of her nose, dressed in a intricately designed, crimson red steampunk-inspired ball gown with a full, layered skirt, ornate brass accents, and a fitted, long-sleeved bodice, walking a sleek, metallic robot dog with glowing blue LED eyes on a worn, brown leather leash down a sun-kissed beach at the golden hour, with the warm, fading light of the setting sun casting a gentle glow on the scene, as the palm trees sway softly in the breeze and the waves crash against the shore in the background, capturing the serene, fantastical atmosphere of a futuristic yet nostalgic era.
Which gives an output like this
You can also combine text and image prompts to get something to your liking, useful because it is quicker and easier to generate images first to check you are getting the desired style before proceeding. For example using the above prompt along with the image below gives this video
There is a more detailed overview of HailuoAI here
KlingAI
- Location: https://app.klingai.com/
- Free Videos: About 8 to start some credits refresh monthly
- Pricing: $0-65 per month
- Features: Images, videos, lip sync, sound, editing, virtual models
- License: Commercial use allowed
I added KlingAI to this article after publishing it because Kling have just become one of my front runners for video generation. Here is a link to the rather nice video with sound which it generated for me from my standard prompt. And here is a still from the video:
Motion is good, image is stable and coherent. The sound is generic but reasonable. KlingAI have a good offering here. Note that the queues for free generation can be very long however.
- Location: https://gemini.google/
- Free images: 20-40 images, daily renewing
- Free Videos: 1 month free trial
- Pricing: £0-19-235
- Features: Negative prompting, image prompting, styles, models, pipelines, additional tools
- License: Commercial use allowed
Veo3 is widely acknowledged as one of the best current video generation models. Google do not offer it on their free tier. Their pricing structure is somewhat opaque but it would appear that you can get five 20 second generations every 4 days for £19 per month and 125 generations per month for £235 mixed in with other features. You can also purchase Veo 3 credits and use them via the API in which case a video costs around £6. They also offer the Veo 2 and Veo 2 fast models with the Veo 2 fast model being extremely economical. While not cheap the results are extremely impressive for realistic video as shown by the examples found here
Leonardo
- Location: https://leonardo.ai/
- Free images: 20-40 images, daily renewing
- Free Videos: 6 daily (Motion 1.0)
- Pricing: $0-10-24-48 per month
- Features: Negative prompting, image prompting, styles, models, pipelines, additional tools
- License: Commercial use allowed
Leonardo creates Video via Google’s Veo3 as well as their own Motion 1.0 and Motion 20.0 models. However since they only grant 150 free credits per day and charge 200 for a Motion 2.0 and 2500 for Veo 3, only Motion 1.0 can be tested for free which only accepts a starting image rather than a starting prompt. Here is an example of a Motion 1.0 clip. This clip was generated based off the following image.
It exhibits good coherence, perspective and fair anatomy, however its animation is largely camera movement and background animation and lacking prompts Motion 1.0 is not easy to tailor as required.
I have covered Leonardo previously here. They are still a strong option for Generative AI
CapCut
- Location: https://www.capcut.com/ai-creator/start and https://pippit.capcut.com/ai-video-generator-v2
- Free Videos: About 3 daily though sometimes you get multiple options for one prompt
- Pricing: $0-11 per month
- Features: Voiceover, styles, models, avatars, additional tools
- License: Commercial use allowed
CapCut offer an unusual video editor that bridges between a desktop application and web based tools. They are a ByteDance Company, but if you do not fancy installing their software you can access their generators online at the addresses above.
Sadly however their video generators do not appear to be general purpose, The Capcut generator produced a panned and scanned slideshow of four images rather than a true video . Meanwhile Pippit got my standard prompt used above badly wrong and produced a series of abstract or talking head narrations rather than the sort of video I was seeking. An interesting product but it does not seem ready for prime time yet. Apart from anything else it appears to mostly offer portrait format video. On the positive side it does seem to be good at voice generation but that was not what I was testing here. Here is a screenshot from one of the strange videos it created.
