sanchitmonga22

6 ms·

1. 31.▲
   by sanchitmonga22 7mo ago

   We built MetalRT from scratch in 48 hours: pure C++ to Metal, no abstractions, no compromises. Result is the fastest decode performance available today on Apple Silicon. 658 tokens per second on Qwen3-0.6B (4-bit) using a single M4 Max. We
2. 32.▲
   Show HN: On-device browser agent (Qwen) running locally in Chrome(github.com)

   19 points by sanchitmonga22 9mo ago|3 comments
3. 33.▲
   Runanywhere – Make every CPU and GPU count(github.com)

   5 points by sanchitmonga22 1y ago|2 comments
4. 34.▲
   by sanchitmonga22 1y ago

   Demo: https://www.youtube.com/watch?v=GG100ijJHl4 Testflight for demo app: https://testflight.apple.com/join/xc4HVVJE Website: https://www.runanywhere.ai/ Follow us for more updates: h
5. 35.▲
   by sanchitmonga22 2y ago

   There's another tool I like to use: Repoprompt, that makes it easier for copying coding files for context.
6. 36.▲
   PrependAI – Solving the Data Ingestion Nightmare for AI Agents(prepend.dev)

   4 points by sanchitmonga22 2y ago|1 comments
7. 37.▲
   by sanchitmonga22 2y ago

   If you're building AI agents, custom connectors for every data source (Slack, GitHub, Google Docs, etc.) are a massive time sink. Prepend unifies data ingestion and lets consumers securely and seamlessly share their knowledge—all throu
8. 38.▲
   by sanchitmonga22 2y ago

   This is awesome!