tattle-made / DAU

MCA Tipline for Deepfakes
GNU General Public License v3.0
6 stars 0 forks source link

Build a proof of concept with feature parity with current BSP for feasibility analysis #130

Closed dennyabrain closed 3 months ago

dennyabrain commented 3 months ago

Work has begun on this repository - https://github.com/tattle-made/kabootar/ Upon pushes to the repository, new version of the package and its documentation is released on https://hexdocs.pm/kabootar/

dennyabrain commented 3 months ago

Built a proof of concept chat bot to see what its like to build directly with the whatsapp cloud API. Some screenshots from the bot

Screenshot 2024-08-09 at 15-18-31 (113) WhatsApp Screenshot 2024-08-09 at 15-18-07 (113) WhatsApp Screenshot 2024-08-09 at 15-17-51 (113) WhatsApp Screenshot 2024-08-09 at 15-17-39 (113) WhatsApp

What you see here is a simple chat bot responding to user's messages with static and interactive messages.

By doing this, I was able to identify 4 verticals, we'd have to work on if we were to build our own message deliveyr solution for Whatsapp.

1. Account Management

2. API Integration

3. Low/No code Journeys

I built a proof of concept for both of these. Created an easy to configure Journey module and customizable blocks in Elixir to evaluate the feasibility of building low/no code chat bot builder. Did similar exploration in UI using low level libraries from d3.js. For the scope of our current requirements, building a low/no code system seems like an overkill, but I felt this exploration was useful to see how feasible it would be to build this if there's demand for it.

4. Infrastructure Management

While we have figured out how to auto-scale given known surge events, handling unexpected surge events would be challenging for this project alone. This project alone does not have the messaging volumes to justify keeping some servers hot 24/7. So just for that reason, it might not be worthwhile to spin up our own message delivery service.

Requirments

What we will require to host