@zarazhangrui As an indie developer, I feel this deeply. Building the product is hard, but explaining what it does, why it matters, and getting people to care can be even harder.
Builders: do you also overthink product names?
I lost an entire afternoon naming a privacy-first, local multimodal file search + smart organizer.
Final pick: Lumora.
Which one would you choose—and how do you name your products?
Other finalists: Mimir, Locus, Zent, Koru, Fathom, Prism, Vesper, Sanctum, Trove, SemVault
#BuildInPublic
Met a friend today in Beihai, Guangxi 🇨🇳She quit OPPO in 2023.
In 3 years she tried 3 “escape routes”:
- content
- ecommerce
- indie products
What money path have you tried (or want to try)?
What surprised me: she only started ecommerce in 2024… in a brutal red ocean… and still made decent money.
Seems like most ex-corporate internet folks end up here:
- Content = sell attention
Low barrier. High consistency tax.
- Ecommerce = connect supply & demand
Margins come from info + service gaps. Needs ops + CS + taste.
- Indie products = create value, sell directly
Hard upfront. Cheap to scale. Compounds like crazy.
That’s the mission: Local, Multimodal, Private.
I'm deep in the code right now, fighting with local inference speeds.
Question for the builders: What’s the one completely offline AI tool you actually use daily?
I need some inspiration (and proof it’s possible).
Building a fully offline AI app is 10x harder than an API wrapper.
No cloud. No recurring server costs. But the engineering challenges? Brutal. 💀
I'm building a privacy-first, local semantic search tool for your desktop.
Because some things shouldn't leave your computer.
I stopped and asked myself:
"Do I actually own my data in the AI era?"
Every prompt, every file uploaded to the cloud... it feels like we gave up privacy for convenience.
We need a middle ground. Smart AI. Zero data leaks.
Building SemFinder: a privacy-first, local semantic search + file management tool.
Today I shipped a semantic relationship graph for your files.
Like Obsidian Graph… but auto-generated from file content (no manual links, no “perfect folder” life).
What should I add next: smart rename, auto tags, or better search UI?
Building SemFinder: a privacy-first, local semantic search + smart file manager.
SemFinder will:
- semantic search by meaning (text + images, later more)
- auto-tag files
- show a topic/similarity graph (Obsidian-style)
- auto-rename messy files
Some people can “just” stay consistent.
Sleep early. Read daily. Train. Ship. Post.
Even with zero feedback.
For me, consistency costs willpower.
BIP friends: what keeps you going long-term—
willpower, personality, or feedback?
#BuildInPublic
@BradAI Makes sense.
My “systems” are more like:
- GTD for tasks/time
- Second Brain for knowledge
But honestly, maintaining them still takes willpower.
Curious: what does your system + environment look like in practice?
And how do you maintain it without it becoming another job?
Build in Public is about sharing the "why" before the "what."
If you’re stuck on what to build next, start with the data.
Would you like the specific prompt I used? Let me know below! 👇
Stop guessing what to build. 🛑
I’ve been refining a workflow to find high-value micro-SaaS niches using AI + Semrush.
Here’s the exact prompt logic I use to turn "seed keywords" into actual business opportunities. 🧵
The final step is Semrush validation.
Looking for that "Sweet Spot":
✅ Low Keyword Difficulty (KD%)
✅ Decent Volume
✅ Clear Intent (Comparison or Tool-seeking)
Look at the screenshot—"TickTick vs Todoist" or "Eisenhower matrix app" are perfect entry points.