Hello all,
I am working on an audio description service, and looking for testers. If you would like to try it out with some of your videos, please contact me. The current version can be found here: https://piccybot.com/audiodescribe/
The final product will offer a limited free tier as well as a pro tier. Anyone can try out the free tier already.
Thanks,
Martijn
Comments
Why don’t you implement this in the app?
Why don’t you implement this timestamp feature in video descriptions in the app? I already pay for PiccyBot, and now I’m going to have to pay for another service as well? This makes things complicated. And adding timestamps to video descriptions has always been something the community has strongly requested for PiccyBot.
For these and other reasons, I’m seriously considering canceling my PiccyBot subscription.
I'd love to test this!!
I signed in and tried to upload a video but got the following error.
ERROR: Could not reach backend. Check tunnel is running.
Not sure if I need to be activated to test, or how it works. But this is a really cool idea!!
Signed in, like it will it be an app?
Hi.
tried it.
yours seems to work better than viddyscribe.
Because viddyscribe describes scenes backward or just in a strange order lots of the time.
yours seems to do it coherently.
the only thing is
do you think it will be in the pixybot app?
I like the service and would pay to help you advance it further.
but I was hoping it could be either a stand alone app or a feature of the pixybot one.
I tried making my own audio description
just a personal one for myself.
but I tried vibe coding it with chat gpt because I can't code well at all.
I ran out of api creddits and tokans.
Is gemini more generous with theres?
.
I'm looking forward to when we figure out how to make a localised model that is a lot easier to use.
Separate service
Guilherme, I am looking at this as a separate service since it's simply quite a bit more expensive to run and I don't want to raise subscription fees for PiccyBot either. Same issue with the Live AI service actually, it works but I can't afford to add it to the current PiccyBot features. PiccyBot will continue to be improved, with a focus on social media descriptions.
Brooke, please keep trying, there are a bit more people accessing the service than I expected, I will add server capacity.
Still no luck
I've tried again and gotten the same error.
Bring my own gemini API key to extend it's use
Hello! Why can't I enter my own gemini api key to use the app even more freely, if I already pay for gemini?
Got it to work once!
I first used ScribeMe's video description feature, and then this tool. Both are amazing, but this one picked up on one detail ScribeMe didn't. It's a video of loads of hummingbirds, and this one mentioned a flock of hummingbirds, while ScribeMe said there were just 2. I know that's down to AI, but I really liked the accuracy. When I tried a second video, I got the usual error. I'd be happy to pay for this once it's more stable.
question
hi, this sounds cool. question: what is the difference between the free vs payed tear?
Pro vs Free
The subscription gives more processing time, direct YouTube sharing, videos upto 30 minutes length and no watermark.
Updated
Hi guys, I fixed a few issues, mainly increasing support for mobile browsers and improved performance. Thanks all for the great support so far!
still doesnt work for me.
when trying to process a 3 minit video, I get — ERROR: Could not reach backend. Check tunnel is running. Also could a progress bar or upload progress be added to know if its even moving? When i tryed it last thirsday, It was on uploading for over 4 hours, and my net is slow, but not that slow.
any updates?
Still broken. iether on phone or pc, i get the same error now.
Backend fix
Ernesto, the backend error issue should be fixed now. Could you try again?
just bumping this a bit.
Do people use the service? I like it better than scribe me, but the free limits are so small its not worth converting my meta glasses videos lol. The limits are even smaller than using free API with obny describer, but I hope when costs settle a bit the dev maybe comes up with a better free option. I'd pay if only to support the piccy bot lable, because we need to keep developers that listen to constructive feedback around, but being broke is no fun currently lol. Also the atempts at making sure we have descriptions of content without risttrictions to the best of allowed AI within the piccy bot app is verry, verry commendable. Also I do have a suggestion. Basically is there a way to add a verbosity level? For example the last video I transcribed was my first time inside a Waymo. My wife and I were talking a lot, and the car itself was talking. So the descriptions seemed to pause while we were talking about things that didn't matter to the video's context. So basically is there a way to Have the app describe even if There is talking in the video? For example, I want to convert videos of me going to this St Jude Three Mile run slash walk that i participated in, But there's a lot of sounds and a lot of talking and a lot of audio, And I do not want to waste my free usage On the Describer staying quiet like it did on my waymo video. It described a few but not much.