AI/ML●●●Banger
Llama CPU Benchmarks
Proves speculative decoding slows down 4B models on 4-core CPUs despite marketing claims.
Big BrainDark Horse
muthuishere
201mo ago
Back to browse
I replaced Whisper with Parakeet TDT on a $55/month CPU server
by TheBuciyo·Jul 9, 2026·2 points·0 comments
AI Analysis
●MidBig Brain
Solid CPU benchmarking but this is a blog post, not a new product launch.
Strengths
- •Int8 ONNX quantization fits 600M param model in 2GB container—practical engineering.
- •Tests accented English, Spanish, Portuguese with real word error rate measurements.
Weaknesses
- •Show HN submission is a blog post about existing service optimization, not new tool.
- •VidClean itself is an established transcription service—this is incremental improvement.
Category
Target Audience
Developers running transcription services on budget hardware
Similar To
faster-whisper · NVIDIA NeMo
Similar Projects
Productivity●●Solid
Local meeting transcription that writes to your Obsidian vault
Plain markdown output feeds Obsidian agent pipelines unlike Otter or Fireflies.
Niche GemSolve My Problem
grembleio
323mo ago
AI/ML●●Solid
AI/ML benchmark for local LLM inference and XGBoost training on GPU/CPU
One-command benchmark suite comparing Ollama and XGBoost performance with a shared Streamlit dashboard.
Solve My ProblemNiche Gem
albedan
201mo ago
AI/ML●●●Banger
MacParakeet – Local voice dictation and transcription for Mac (GPL-3.0)
155x realtime transcription on Neural Engine beats Wispr Flow at zero cost.
Solve My ProblemSlickDark Horse
moona3k
103mo ago
Productivity●●Solid
Local-first dictation for macOS – Parakeet TDT, zero cloud calls
Free local dictation when Superwhisper and Wispr Flow already charge monthly.
CozySolve My Problem
newtechwiz
103mo ago
Productivity●●●Banger
OpenGranola – meeting copilot that searches your notes in real time
Searches your notes during calls while Otter.ai just transcribes and waits.
Solve My ProblemSlickDark Horse
yazinsai
103mo ago