AFriB

Inspiration

Other AI platforms like Gemini, ChatGPT and Claude are all tuned to respond to our prompts and questions but they are not able to help with the common things we do, like

asking AI to remind me to come back to a conversation,
asking AI to suggest groceries to buy with my budget and once it generates that, add it to my task so I can remember to pick up the groceries after work.
getting UI feedbacks during my conversations which helps clarify what I need without my conversations feeling like I'm still conversing with a Bot in 2026.
getting updates about the Stocks I'm interested in daily without breaking my bank.
Getting job updates without having to setup a $500 to $1000 open claw or mini server at will.
getting AI to understand what I did last week, and if I've improved now.
Getting AI to check up on me based on our previous conversations. and many more.

All of these are the current limitations of our current AI platforms, most of them can do some of this, but they are mostly beneficial to users with technical abilities and with thousands of dollars to spend. AFriB wants to change this, AFriB wants to make all these and many more possible from one interface with many more features being planned and released weekly. AI should truly be intelligent, and that means AI should replace our digital efforts so we can focus on real-life activities and personal connections. This is what AFriB wants to achieve.

What it does

AFriB can do everything the current AI platforms can do, except be a coding agent (because it wasn't programmed to do that), AFriB main focus is not developers, but the everyday users, those who have seen the Sci-fi movies of what AI can do, but haven't experienced it yet. We are bringing that sci-fi experience to our users with AFriB.

How we built it

We built it by leveraging existing LLMs and building our own rich harness on top of the LLMs, we tuned every response to be significant to the user so that every conversation feels like a fulfilled activity, and not more questions.

Challenges we ran into

We ran into several challenges.

Ensuring that every response is accurate and meant to solve the user's request.
Getting the platform to run for our users as cheap as possible.
Managing the different agents which handles different tasks since we are targeting both technical and non-technical users.
The UI was a big challenge, because we needed to ensure that we provide a good interface but also accompany that with a better User Experience, and that required multiple iterations before settling on something we believe works for all users and for all use-cases.

Accomplishments that we're proud of

AFriB can do what other AI platforms can currently do, and more, we want AFriB to be an everyday companion for anything you practically do. Think like a mini-AGI. That's what AFriB is, and we are proud of how we have accomplished this.

What we learned

We have learnt that building Artificial General Intelligence is not only about the models, the models already understand almost everything we humans understand and probably more, but building the next frontier AI agents, products or Friends requires building another layer of personality on top of the LLMs that understand what it means be to a human and what it means to he a helper, a Friend and a companion to humans. That's where we can truly see the impact of AI.