AI tool comparison
Ghost Pepper vs Sup AI
Which one should you ship with? Here is the side-by-side panel verdict, pricing read, reviewer split, and community vote comparison.
Productivity
Ghost Pepper
100% on-device speech-to-text and meeting transcription for Mac — zero cloud
75%
Panel ship
—
Community
Free
Entry
Ghost Pepper is a macOS menu bar app that runs Whisper-based speech recognition and meeting transcription entirely on-device via Apple Silicon — no internet connection required, no audio leaving your machine. Hold Control to dictate into any text field; it transcribes and pastes the result in seconds. For meetings, it records calls and generates full transcripts, notes, and AI summaries saved as local markdown files. The app supports multiple model sizes from a 75MB fast model to a 1.4GB multilingual option covering 25+ languages. A local LLM layer (Qwen 3.5 variants) strips filler words and self-corrections from transcripts. The developer published a privacy audit confirming zero cloud API calls, tracking SDKs, or telemetry in the core functionality — an unusual level of transparency in this space. Built on WhisperKit and LLM.swift, Ghost Pepper requires macOS 14.0+ and Apple Silicon. It launched on Product Hunt today reaching #4 daily. For anyone running sensitive client calls, legal conversations, or just unwilling to feed voice data to cloud services, this fills a genuine gap that ElevenLabs, Otter.ai, and Whisper API don't touch.
AI Productivity
Sup AI
Runs 339 LLMs in parallel and downweights the hallucinating ones.
50%
Panel ship
—
Community
Free
Entry
Sup AI is an ensemble AI assistant that runs your query through 339 language models simultaneously, measures per-segment confidence across all responses, and synthesizes a final answer that amplifies agreement and suppresses likely hallucinations. The team claims a 52.15% score on Humanity's Last Exam (HLE) — 7.41 percentage points above the single best model — which, if verified, would make it the highest-scoring system on the benchmark to date. The underlying mechanism works like an LLM panel: each model votes on sub-claims within the response, confidence is estimated by agreement density, and the final output surfaces high-confidence segments while flagging uncertain ones. It's designed to reduce hallucination rate on factual tasks, not improve reasoning per se — the models in the ensemble aren't doing collaborative chain-of-thought, they're voting on outputs. Sup AI was built by Ken Mueller (Stanford, CEO) and Scott Mueller (AI Research Scientist) and launched on Product Hunt today. Pricing starts with $10 in free credits, no auto-charge, with a credit card required to start. The HLE benchmark claim is the headline and will face scrutiny — if verified, this is a meaningful research result. If it's cherry-picked, it's still a usable product with a differentiated architecture.
Reviewer scorecard
“WhisperKit on Apple Silicon has gotten fast enough that local transcription is genuinely competitive with cloud services in latency. The Control-to-dictate UX is exactly right — no separate app to open. The privacy audit documentation is a rare and welcome move for an open-source tool.”
“The HLE claim needs independent verification, but the underlying ensemble approach is architecturally sound for factual Q&A tasks. Running 339 models is expensive — pricing will be the gating factor for production use. The $10 free credit is a fair trial.”
“Apple Silicon only is a real limitation — no Intel Mac support, no Windows, no Linux. The meeting transcription accuracy will lag behind purpose-built cloud services like Otter or Fireflies that have years of model tuning. And the 1-7 second cleanup latency adds up in fast-paced conversations.”
“Extraordinary claims require extraordinary evidence. A 7.41 point jump on HLE via ensembling — without publishing methodology — smells like benchmark gaming. The latency of running 339 models in parallel is also a real concern for anything other than async research tasks.”
“This is the inevitable direction: voice AI moving entirely on-device as hardware catches up to the task. Ghost Pepper is the leading edge of a shift where sending voice to the cloud will feel as strange as sending passwords to cloud storage does today. Apple's Neural Engine investment is paying dividends here.”
“Model ensembling is an underexplored direction in the race to reduce hallucination. If Sup AI's approach scales, it could be more durable than fine-tuning individual models — you get the wisdom of the crowd across model families, training data, and architectures simultaneously.”
“The name is perfect — spicy, memorable, evokes both heat and ghostly invisibility (no data leaving). Menu bar apps with zero UI overhead are the ideal form factor for voice tools. The markdown output for meeting notes plugs straight into any PKM workflow.”
“For creative work, ensemble outputs tend to regress toward the mean — you get the most-agreed-upon version of something, which is usually the least interesting version. This is a tool for factual accuracy, not creativity. I'd stick with a single strong model for writing.”
Weekly AI Tool Verdicts
Get the next comparison in your inbox
New AI tools ship daily. We compare them before you waste an afternoon.