What do you like best about fal?
It houses all the leading AI models within one platform, and you can access the best of all models. You have the choice to select which one to use depending on the occasion, your final objective, and the kind of project you are working on. I particularly love the recently added custom modification of the MiniMax H3 model, which they have titled as H3Max. It has a fast version that generates video at a very fast rate without compromising on the output quality of the video clips. I also loved the fact that once you buy certain credits, which you can spend on any kind of AI model, those credits do not expire for 1 year. If you have bought $20 worth of credits, you can reserve them for a right time when you can use them, and that way, the pricing is very reasonable. Review collected by and hosted on G2.com.
What do you dislike about fal?
Some of the aspects that I dislike about fal are that, when you go by the system of buying credits per use, that does not automatically imply that the services are cheap. Sometimes the cost can really go high.
You need to be extremely careful about how many credits are consumed per image you generate with models such as Nano Banana or GPT Image 2.5 (which are the latest models), or when you use video generators such as Seedance 2.5, Kling, and Google Veo. When you use such kinds of video models, you have to be extremely careful to keep track of exactly how many credits are consumed per second when you generate using different kinds of video resolutions (such as 720p, 1080p, or 4K) and the duration of the clips. You have to be constantly vigilant to monitor such kinds of credit consumption so that you do not exhaust your credits even before achieving your objective.
Another huge disadvantage is that you do not get to know when the credits expire. There are no alerts, no prior notifications, and you have to keep monitoring on your own the exact expiry date. Sometimes it is in 90 days, sometimes it is in 365 days, and if you fail to keep track of such expiry dates, then you might miss out on the deadlines and you might waste credits that you never utilized. It is always advisable to purchase credits in proportion to the amount of work or usage you will be doing. It is better to be practically aware of that.
Another problem is the integration or compatibility with certain AI tools. Unlike other competitor tools, where you can have a very attractive UI interface, in here, the settings for several of the tools do not have user-friendly options for you to tweak the settings. You have to be an advanced user most of the time to get the best out of such an interface. They haven't put in the effort to make the interface user-friendly. On top of that, if you want to learn how to use certain kinds of tools, there is no help documentation corresponding to such tools, and their support expects you to find such guidance elsewhere outside the website. Unlike other websites where you can use several AI tools together, in fal, you do not get a similar type of interface. It's more inclined towards coders or technical experts.
When it comes to integrations, some of the leading prominent models are still not exactly optimized for the fal interface. If you use custom Lora with models such as Flux, in such cases, it can be very painful for you to master the advanced parameters, which are not user-friendly in the interface. When you seek support, you don't get it too fast. Sometimes the responses are too slow, and you don't get the guidance or the onboarding that is required in crucial moments where you are confused or clueless. Review collected by and hosted on G2.com.