Overview of My Submission
Chattergram is a chat application POC that translates your voice messages to other people into written text. I've built it during a previous hackathon & it turned one of my favourite side projects. I didn't score a price, but learned a lot about Typescript, Vue 3 and Speech To Text functionalities.
Now, I'm proud to present its enhanced successor:
Introducing Chattergram - powered by the Redis Stack
Before, the project didn't feel complete. In retrospective, I came to think that some choices for the architecture were suboptimal at best:
- I was using no serverside caching
- Some of the backend components were tightly coupled (e.g. the Chat - handler had a hard dependency on the Websocket Server)
- PostgreSQL felt like an inappropriate choice for such a trivial data structure. I've spent a lot of time to properly setup the data structure & profit from TS intellisense.
In the following picture, you can see the pre-Redis application structure:
I'm happy to say that using Redis allowed me to address all these problems fairly straightforward:
- Redis now acts as my frontend database
- It syncs Chat and User entries to - and from MongoDB
- Redis' pub/sub functionality allowed me to de-couple my Chat handler & the Websocket server
- While MongoDB does not feel easier to navigate, I found it to be an easier & faster persistent database to get up and running quickly.
Below is the new application structure. Almost a replica of how I've had it planned out (writing base64 encoded files didn't turn out to be a great idea):
Using Redis also boosted performance quite a bit. You can find some simple benchmark results in the Github Repos' readme
Chattergram uses a frontend component built with Vue3. It also offers an OpenAPI specification to help developers understand its functionality.
You can find the OpenAPI specification under
http://localhost:8080/api/docsonce the app is running locally.
Submission Category:
MEAN/MERN Mavericks
Actually, MFVN - Mavericks (MongoDB, Fastify, Vue, Node)
[Optional: Video Explainer of My Project]
Language Used
- Node.js + Typescript
Link to Code
tq-bit
/
chattergram-redis
A fullstack chat application that transforms speech messages to text. Built for the Deepgram Hackathon 2022
Chattergram + Redis
A chat application that transcribes voice messages to text - powered by the Redis Stack
Chattergram runs on two separate applications - a frontend and a backend. You need to create an account to access the chat. All of your data will be stored on a local Redis (& MongoDB) instance.
Once signed up, you can start chatting with other users. If you decide to send them a voice message, Deepgram will try and transcribe the message for you and your sender. You can still listen to the recorded message if you like.
I've created Chattergram-Redis as a POC project for the Redis Hackathon on dev.to. Please do not use it in a productive environment
How it works
Video explanation
You can find a video version of my submission on Youtube:
How the data is stored:
- This app uses two primary data types:
Userand…
Additional Resources / Info
- Chattergram original (=legacy) readme file
- Deepgram console login
- Apache Jmeter performance test tooling
Collaborators
Solo - project
- Check out Redis OM, client libraries for working with Redis as a multi-model database.
- Use RedisInsight to visualize your data in Redis.
- Sign up for a free Redis database.
Top comments (2)
I'm happy you liked it :-) thanks for giving it a try.
This is really cool! Great job.