Hybrid vector + keyword search across 4M Wikipedia articles. Built in 4 months by a self-taught dev on recycled crypto mining hardware. 128ms queries, 104GB VRAM across 4 GPUs. Part 2 of 5 launches this week.
Hey Product Hunt! 👋
The Journey:
4 months ago, I couldn't run a terminal command. Today I'm launching my second vector search API.
The Stack:
1x RTX 5090 + 3x RTX 3090s (104GB VRAM)
Dell R740XD with 315GB RAM (found free)
Recycled crypto mining infrastructure
Total build cost: ~$5k
The Results:
4 million Wikipedia articles vectorized
128-213ms hybrid search queries
Vector + keyword fusion for best results
Free public API: https://wiki.built-simple.ai
This Week's Launch Schedule:
✅ Monday: Stack Overflow (live at fixit.built-simple.ai)
🔴 4.8M Wikipedia articles
⏳ 2.7M ArXiv papers
⏳ 4.5M PubMed medical articles
⏳ 9.2M legal documents
Why?
Proving we can vectorize ANY domain at scale. Next step: Custom vectorization for your data.
Want early access to turn your knowledge base into searchable vectors? Drop a comment 👇
Report
No reviews yetBe the first to leave a review for Wiki Vault