HACKER Q&A
📣 HNtemp

How to build group chat app in 2020?


I am going to build a group chat app (for both iOS/android). I already have an MVP but for next version I want to make sure it can scale smoothly to hundreds of thousands of users (be optimistic!)

Need just standard features: Multiple groups/channels up to a couple thousand people each, quote/reply, pictures, direct messages, profiles, backend panel.

Not technical, but I am somewhat knowledge, worked on basic tutorials but will not be able to build anything robust by myself.

I would very much like to use as much pre build stuff as possible so I focus on the unique selling proposition (why people will actually use) and connect that to the core group chat.

What is the problem? Too much info/options. What framework, languages, database, messaging protocol to pick? From free SDKs, to paid stuff, pre made scripts (though none quality?). Lots of options, no clarity. All suggestions and thoughts welcome, especially specifics. Thanks.

  👤 core-questions Accepted Answer ✓
Do we really need another chat app? Wouldn't it maybe be better to capitalize on the existence of an open network that is already around, and build a better client for it that kicks the pants off of whatever's already out there?

> will not be able to build anything robust by myself.

Then who's going to build it? Are you hiring a team to do this? Don't underestimate what's involved here to get something working and keep it running. At the very least, if you don't go with a decentralized model, you're in for expensive server bills and considerable ongoing maintenance to keep up with security patches, etc.

How will you monetize this to pay for all that infrastructure?

> Too much info/options

There are already too many options in this space. Why would anyone choose your client when Signal, Telegram, Riot/Matrix, etc. are all free and decent to use? What is your unique value-add?

Will you be writing a client for multiple platforms? Desktop? Web browser? Don't be fooled into thinking you can build one client that works everywhere, you're up against native apps with native integrations that make sharing and interoperation easy - features that take whole teams to get right, much of the time.

> All suggestions and thoughts welcome, especially specifics. Thanks.

Don't do this unless you have a business plan good enough that someone will lend you $1MM to get it off the ground. Seriously. You're talking about entering a massively crowded marketplace, and you're clearly out of your element.

If you must do something, make a paid Riot/Matrix client that doesn't suck as much as the stock one does, and steal users from an existing network.