SnackShell

Scammers please Leave me alone 😭

From commits to commitments. Wishing you the best. May it age gracefully. ❤️

I want to share something early. SelamGPT is being built as a unified home for many leading AI models...including Claude 4.5 Sonnet and other Claude models, GPT-5.2 and additional GPT models, Gemini 3 Pro, Kimi K2, MiniMax M2, GLM-4.7, and other 100+ models. Beyond text models, Selam will also provide image generation, translation, text-to-speech, search, and live AI agents. all accessible through a single Selam API with Openai compatible sdk. The goal is simplicity: one platform, one integration, many capabilities. If everything goes as planned, we’re aiming for an initial launch in the first month of 2026. More details will come when the time is right.

New for Developers 🚀 Developer accounts are now live on AfuChat. Verified dev profiles get trusted status, clearer identity, and access to tools to build and improve the platform. If you’re building, experimenting, or contributing—this is for you. DevSDK | Contribute | Dev profile

SelamGPT Progress Update We know it’s been a while since our last update, and we truly appreciate your patience. Behind the scenes, we’ve been tackling some major challenges in our backend infrastructure. Our goal has always been clear: to keep SelamGPT free and accessible for everyone. Finding reliable free providers and making everything work smoothly has taken time, effort, and careful engineering. At the same time, we’re building an all in one SelamGPT SDK. bringing together text, text-to-speech, image generation, translation, and more into a single, unified experience. We’re now actively working on the frontend, shaping the product you’ll soon interact with. Progress is steady, and we’re getting closer every day. Thank you for believing in the vision. Stay patient, stay tuned — the journey is moving forward. 🚀 #ProgressUpdate 🖥 For Early access: Make sure you’re on the waitlist: https://selamgpt.com

Gemini don't know a tool called "Nano banana Pro." 🤬

Merry Christmas 🎄🎅 #telemetrio2025

Afuchat by @amkaweesi is now Open Source! Contributors can explore the code, submit improvements, and help build new features. Production remains fully protected and secrets are safe. Check it out here: https://github.com/amkaweesi/afuchatv1

Snackshell on AfuChat Check out Snackshell's profile on AfuChat https://afuchat.com/snackshell

Make sure to select the default monthly plan!

Get Lovable 2 month pro for $0. Use this promo code: NEXTPLAY-LOV-25 https://lovable.dev/

GM. Who's online right now?

Who's online right now?

Here is a short video how Ra'yee works.

Now, after some modification: - we get rid of its ugly shape and make it look like a glass - we added a nose holder, made it flexible and comfortable to wear. - we redesigned and decreased it's thickness by 3x - it had a bad battery capacity (250mah), now we changed it to 1500 mah. - we modified the camera to shot ultra wide as much as possible.

Now, after some modification: - we get rid of its ugly shape and make it look like a glass - we added a nose holder, made it flexible and comfortable to wear. - we redesigned and decreased it's thickness by 3x - it had a bad battery capacity (250mah), now we changed it to 1500 mah. - we modified the camera to shot ultra wide as much as possible.

Now, after some modification: - we get rid of its ugly shape and make it look like a glass - we added a nose holder, made it flexible and comfortable to wear. - we redesigned and decreased it's thickness by 3x - it had a bad battery capacity (250mah), now we changed it to 1500 mah. - we modified the camera to shot ultra wide as much as possible.

This is a smart glass for the blind community. It captures an image with a tactical button attached to an esp32 cam from the glass (wireless) and captures the image from the streaming video in the Mobile I'd built, then it sends it to my backend (Meta Llama 4 with groq api) and translates the response to the selected language in settings, then it convertes the text to speech (tts) and sends guided audio response for the blind person through the wireless blutooth earpod following the given system instructions in my backend server. I could have used Gemini models (since they're good in Ethiopian languages, which doesn't ask for translation) but now days they have decreased the free tier limit, so it hits the limit too fast...That's why am using a translator. The glass also have an Ultrasonic Distance Sensor + Buzzer for quick guidance. It sends sound frequencies between 20 Hz and 20 kHz and the beeping speed increases or decreases depending on how close an object is, especially within 1 meter.