Show HN: SubSparks – I built an AI to turn Reddit pain points into SaaS ideas

Hey HN,

I'm Victor, a solo dev, and I'm excited to share a project I've been passionately working on: SubSparks (https://www.subsparks.com).

The Problem I Was Trying to Solve:

Like many here, I love the idea of building MicroSaaS products. The biggest hurdle for me has always been the initial step: finding a genuinely validated problem to solve. Hours spent brainstorming, browsing forums, and still feeling like I was shooting in the dark. I noticed Reddit is a goldmine of people discussing their frustrations and needs, but manually sifting through it for real opportunities is incredibly time-consuming and often hit-or-miss.

What is SubSparks and How It Works:

SubSparks is a platform designed to automate this discovery process. At its core is an "Insight Engine" I built in Python. Here's a high-level look at how it operates:

Reddit Monitoring: The Python engine, based on user (team) configurations (target subreddits, keywords), actively scans Reddit for relevant discussions using PRAW. AI-Powered Pain Point Identification: I'm leveraging Google's Gemini API (specifically the Flash model for a good balance of capability and cost) with carefully crafted prompts. The AI analyzes the collected content to pinpoint specific "pain points," frustrations, or unmet needs that could potentially be addressed by a software solution. It's not just keyword matching; it tries to understand the context and intent. MicroSaaS Idea Generation: For each significant pain point identified, the system (again, using Gemini with different prompts) generates potential MicroSaaS ideas. This includes a name, a brief description, target audience, key features, and even a suggested monetization strategy. The goal is to provide a concrete starting point. Data Storage & Delivery: All these insights (raw posts, identified pains, generated ideas) are stored in a MongoDB instance. The SubSparks SaaS platform (built with Next.js and Postgres for user/team management) then presents this data to the users in a structured way. The Python engine itself runs on a VPS, which I'm managing with EasyPanel. This gives me the flexibility to run and update the scraping/processing logic independently of the main SaaS application.

Why I Built This & What's Different:

I wanted to move beyond just "idea generators" that spit out random concepts. SubSparks aims to ground every idea in a documented pain point from a real community. The link back to the original Reddit discussion is always provided, so users can dive deeper into the context and validate the pain themselves. The "value" is in automating the initial grunt work of sifting through vast amounts of text and connecting dots that might otherwise be missed.

Tech Stack (Briefly):

Insight Engine (Backend): Python (PRAW, Requests, PyMongo), Google Gemini API. SaaS Platform (Frontend/User Management API): Next.js, TypeScript, Drizzle ORM, PostgreSQL. Databases: MongoDB (for Reddit data & insights), PostgreSQL (for SaaS platform user/team data). Engine Hosting: VPS with EasyPanel. Current Stage & What You Can Try:

SubSparks is live, and you can configure your own "monitors" to start discovering pains and ideas in niches you're interested in. There's a https://www.subsparks.com so you can test it out.

I'd Love Your Feedback!

This is very much a work in progress, and the HN community's perspective would be invaluable:

What are your thoughts on using Reddit + AI for SaaS idea generation? How do you currently find and validate your project ideas? What are your biggest pain points in that process? Are there specific features you'd find useful in a tool like this? Any suggestions on how the "pain score" or "idea validation" aspects could be improved? Curious about any specific technical challenges I faced with the Python engine or Gemini integration? Happy to share more. Thanks for checking it out! I'll be around to answer any questions.

Victor