Typewriting AI

• 

  Feedback, made human

Inspiration

In the real world, when a customer has something to say, they talk to you. Like a human.

They don't just click stars or fill out rigid forms. They speak. They vent. They ask. And sometimes, they don’t even know exactly what they want — they’re looking for clarity, not just a survey.

They want to have a conversation — not fill in boxes.

But most digital feedback systems don’t allow that. You send a form, they tick a few options, and maybe drop a quick comment. That’s it. Then they leave, often unheard.

Even if you're great at solving problems, the tools don't give you the chance to ask the right questions or explain your service properly. There's no real connection, no chance to adapt in real time.

What it does

We're building a new kind of feedback tool — one that listens like a human, asks intelligent follow-ups, explains your product or service clearly, and turns feedback into a two-way conversation.

It helps users feel heard, understood, and informed — so they don't leave out of frustration or confusion.

How I built it

We built on AWS infrastructure using different services in the cloud connected each other to perform in the most efficient manner without spending a lot of cost.

Challenges I ran into

LLM inferencing and vector databases come with big expenses as well as keeping the embedding properly distributed is important for the efficiency. At the same time it is also important to save user's data from being used for machine learning and not mixed up. We had to design the whole architecture keeping all these things in mind. Also, as I didn't have fund to hire anyone, we lacked workforce to build it.

Accomplishments that I'm proud of

We've got featured on Product Hunt and we're humbled to get good responses from users and many of our team members volunteer to make the product better.

What I learned

I learned how to iterate an idea until getting a product market fit. Also learned how to lead with less funding to get maximum output.

What's next for Typewriting AI

We've opened our mission for anyone who wants to contribute to make this product better. So, everyone is welcome to try, give feedback, or provide us with things that help us grow inside and in the market. We're iterating the whole UI and backend actions and update every week to make it perfect.

Built With

• ai
• amazon-web-services
• embedding
• inferencing
• llm
• machine-learning
• mistral
• mysql
• php
• react