I built Seer to solve my own problem: I don't have patience for social media, but I still want to find market opportunities.
Seer monitors Hacker News, GitHub, npm, and DEV.to for signals like:
- "I wish there was..."
- "Looking for alternative to..."
- Trending repos with "help wanted"
- New packages gaining traction
It scores opportunities (0-100) based on engagement, recency, and relevance, so you see the good stuff first.
*Tech:* Go backend, SvelteKit frontend, SQLite. Self-hosted, runs with `docker run` or as a single binary.
*No cloud, no tracking, no Pro tier.* 100% open source.
I built this because I needed a simple way to generate invoices without subscribing to yet another SaaS. Most invoicing tools are either overkill (Freshbooks, QuickBooks) or cloud-only with monthly fees.
Stack: SvelteKit + SQLite + PDFKit. Runs locally, data stays on your machine.
What it does:
- Create invoices with line items, tax calculation, client management
- Generate professional PDFs
- Single SQLite file (easy backup, just copy the file)
- Works offline
What it doesn't do:
- Payment processing (it's just invoicing)
- Accounting (not trying to be QuickBooks)
- Cloud sync
$39 one-time, full source code included. No subscriptions, no accounts, no tracking.
Happy to answer any questions about the stack or implementation.
Mendex here. I'm excited (and a bit nervous) to share a project I've been pouring my heart into: CubicLog.
I've actually never used a traditional logging stack like ELK or Grafana/Loki. When I needed a logging solution, I looked at the market and saw a level of complexity that felt fundamentally opposed to my philosophy of how software should be: simple, self-contained, and respectful of the user's time and resources.
Instead of adopting a tool I didn't believe in, I decided to build the logging tool I wished existed, from first principles.
The result is CubicLog. It's not an "alternative to X"; it's a manifestation of a philosophy:
Single Binary & SQLite: The entire application, including the web UI and database, is a single, ~11MB cross-platform binary. No dependencies, no Docker required.
"Simple by Design, Smart by Default": The core idea. You can send completely unstructured logs, and the built-in pattern matching engine (with over 460 patterns) automatically analyzes the content to derive metadata like severity (`error`, `warning`, `success`), extract the source service, and identify patterns like stack traces, HTTP errors, or security issues, all with zero config.
Beautiful Embedded UI: I believe simple tools deserve a premium user experience. The entire dashboard is a modern, responsive UI powered by Alpine.js and Tailwind from a CDN, embedded directly in the Go binary.
Full Service Management: Includes CLI commands to run as a service (`-stop`, `-status`, `-restart`) using a PID file.
This project was built to solve my own frustrations, born from a long battle with burnout where simplicity became a necessity. It was built in a unique Developer-AI collaboration, which is documented in the repo.
I'm here all day to answer any and all questions. I'd love to hear your thoughts, criticisms, and ideas. The project is fully open-source (MIT license).
I built Seer to solve my own problem: I don't have patience for social media, but I still want to find market opportunities.
Seer monitors Hacker News, GitHub, npm, and DEV.to for signals like: - "I wish there was..." - "Looking for alternative to..." - Trending repos with "help wanted" - New packages gaining traction
It scores opportunities (0-100) based on engagement, recency, and relevance, so you see the good stuff first.
*Tech:* Go backend, SvelteKit frontend, SQLite. Self-hosted, runs with `docker run` or as a single binary.
*No cloud, no tracking, no Pro tier.* 100% open source.
GitHub: https://github.com/mx-seer/seer
Would love feedback on the scoring algorithm and what other sources would be useful.