I keep getting the "Error: Failed to stream response from [Gemini | OpenAi] API. Please try again." - tried valid new keys from both google/openai

Is that with 2.5 Flash? I got that error intermittently with that mode, but the other Gemini models worked fine. I'll investigate

ah yea 2.0 flash is working. 2.5 doesn't and OpenAi 4.0 and mini models dont work either. The error message should probably say to try other models because i was pretty confused

out of technical curiosity, how do you render the video? is the whole thing like a canvas element?

Combination of several rendering techniques.

The idea is as long as you have all the information the user wants in in each layer. Then you can do a lot of things with this understanding.

The major pieces of information you need to collect while working on a 2D plane are these.

- The x,y coords - The width and height - A path to the asset - additional properties like colors and opacity

From having this critical information stored nicely means you can email it to a graphic designer and have them decode it and follow the "map".

You can send it to a client side process and have it interpret the information.

You can send it to a server and have it interpret the information.

You can do a one shot kill and build a system that pieces them together using one approach.

You can send each layer to a different process. You can create a specialized technique for each layer type.

The key is collecting the information. The actual rendering can be done in many ways. FFMPEG, canvas, screen shots, send it to a human lol. Using OS commands, etc.

"send it to a human" lol that made me chuckle!

Thanks for explaining this in detail, Michael! Really great work here!

Haha it's all about communication. Glad you got a chuckle.

They still carry it. And can in turn infect their parents

Hasura has this feature, but unfortunately it seems like its part of their "pro/enterprise" offering and not the OS version. https://hasura.io/hasura-pro

@elithan this looks awesome man. The one thing keeping me in the firebase ecosystem is the firestore rules (basically quick CRUD permissions). Does nhost provide something similar?

Yes. Hasura handles this in their permission system and they do an excellent job. Super easy to get started with.

Together with Nhost's auth system and Hasura, it's easier than Firebase.

Isn't this against facebooks ad policy? If i drive traffic to a look alike app store that doesn't do anything.

Thanks for the feedback! Do you have more information on this for me? I do not see a reason why it should be against their ad policy. I am not aware of any ads that are driving traffic to it. Furthermore, the whole copy of the site clearly states that it is a previewer.

@treblig any advice on usually how much percent increase you see once you move up funnel with fb advertising (optimize)? I'm guessing the cpc you got are based on bottom funnel and highly targeted hence the high cpc. I find really difficult to estimate costs with fb, because it takes me a fairly long time (1-month+) to build the content and ad strategy to lower costs. But I'm wondering at what point do you know if its a bad business idea vs bad ad strategy?

I love this idea, it's kind of like an affiliate network for physical goods. The delivery time is probably the only thing the scares me.

Us too. We're keeping data on sellers and their fill, ship and complete time. Fill time meaning how fast they fill when the order gets to them, ship time meaning how long it takes from the order time until the shipment occurs, and complete time meaning from time of order to time you get notified.

I initially thought this was post for https://github.com/ory/hydra

Or the Emacs library.

a lot of that is water weight, I almost always have to average a weeks worth of weight to get a stable reading on the weight changes. I also, gain like +2lbs if I eat a salty meal the night before.