Spaces:
Running
Need feedback on a application
Hey folks 👋
I’m building an AI-powered app and looking for early feedback from experienced devs here. The website is asteai.com
please paste the link in the browser and try signing up free
Would really appreciate if you could try it out and share honest feedback — what works, what doesn’t, and what you’d improve from ui to ai and also performance and latency
Infra-wise, current setup is:
• AI model → EC2
• Backend → Render
• Frontend → DigitalOcean
Since this is still early-stage, I’m also looking for guidance on scaling this efficiently (cost + performance + architecture).
If you’ve worked with similar setups or scaled AI apps before, your insights would be super valuable 🙌
For an early AI app, I would separate product feedback from operating feedback.
Product side:
- What is the first user action after signup?
- How quickly can a user get one useful result?
- Is the AI output tied to a concrete workflow, or does it feel like a generic chat box?
Infra side:
- log one run id per user request
- capture model latency, backend latency, and frontend render time separately
- track retries/timeouts as first-class metrics
- keep a small eval set from real user failures
- measure cost per successful user task, not only cost per model call
The EC2 / Render / DigitalOcean split can work early, but debugging gets painful unless each AI request has a clean trace across frontend, backend, model call, and final response.