Ghost Pepper – Local hold-to-talk speech-to-text for macOS
I built this because I wanted to see how far I could get with a voice-to-text app that used 100% local models so no data left my computer. I've been using a ton for coding and emails. Experimenting with using it as a voice interface for my other agents too. 100% open-source MIT license, would love feedback, PRs, and ideas on where to take it.
FL score
out of 100
Verdict
high confidence
Competition
7
competitors found, emerging market
Trend
No signal yet
A local, privacy-focused hold-to-talk speech-to-text app for macOS users, with strong potential as a voice interface for AI agents.
The pain
The gap
Build angle
Strengths
- High demand for privacy and local processing, validated by user complaints about cloud competitors.
- An MVP is already built and open-source, reducing initial build risk.
- Strong founder-product fit (builder uses it personally).
- Clear differentiation potential with 'voice interface for agents' for power users/developers.
- Bootstrapped competitors show a viable one-time purchase monetization model.
Risks
- Monetization strategy for an open-source project needs to be carefully defined.
- The 'agents' integration niche might be too small initially, requiring broader appeal.
- Competing with 7 other bootstrapped tools requires strong execution and marketing to stand out.
- Building a sustainable business from an open-source core can be challenging.
Next steps
- Conduct user interviews with macOS power users and developers (especially those using AI agents) to validate demand and willingness to pay for 'agent integration' features.
- Explore potential business models beyond pure open-source: premium features, a polished Mac App Store version, consulting for integrations, or a 'buy me a coffee' model.
- Gather specific feedback on the 'hold-to-talk' interface – is it preferred over continuous dictation by the target audience?
- Research the specific feature gaps within existing local macOS STT apps that the 'agents' integration could address (e.g., custom APIs, webhooks, specific scripting languages).
Fly Labs Method
Is the pain real, is there a gap, is it the right time, can one person build it.
- Problem clarity
- 25
- Solution gap
- 18
- Willingness to pay
- 21
- Buildability
- 13
Solid problem space with clear demand for privacy and local processing, backed by existing competitor pricing. The current solution has an MVP and a unique 'voice interface for agents' angle.
Value Equation
Dream outcome and how likely it feels, against the time and effort it costs.
This idea offers a strong value proposition in a growing market, addressing clear pains with a product already built and potential for differentiation.
One-Person Business
Curiosity pull, identity fit, and a path from free value to paid for a solo creator.
Strong creator fit and clear problem, but monetization strategy for an open-source project needs refinement for solo builder sustainability.
Viral Frameworks
Hook strength, shareability, and how cheaply it can be tested.
Strong value proposition and target audience, but the open-source nature necessitates a clear business model and further validation for paid offerings.
Builder Lens
Evidence the problem exists, timing, defensibility, and a model that fits on a napkin.
Strong validation on real demand, specific user, and future relevance, with good potential for a narrow, shippable wedge.
Why this verdict
Five lenses, one composite. How scoring works
The angle
This weekend
- Conduct user interviews with macOS power users and developers (especially those using AI agents) to validate demand and willingness to pay for 'agent integration' features.
- Explore potential business models beyond pure open-source: premium features, a polished Mac App Store version, consulting for integrations, or a 'buy me a coffee' model.
- Gather specific feedback on the 'hold-to-talk' interface – is it preferred over continuous dictation by the target audience?
- Research the specific feature gaps within existing local macOS STT apps that the 'agents' integration could address (e.g., custom APIs, webhooks, specific scripting languages).
Who is already there, emerging market
- SpeakmacVisit
Speakmac is a macOS app that transcribes your voice into text instantly, privately, and fully offline without draining your CPU, aiming for a minimal and polished user experience.
Pricing: One-time purchase (details not specified, but the description mentions "No subscriptions")
- SuperwhisperVisit
Superwhisper offers accurate, on-device voice-to-text for Mac and iPhone, prioritizing privacy by running OpenAI's Whisper entirely locally.
Pricing: One-time purchase of $150, or $15/month for lifetime access to model updates (as of early 2024 pricing noted in external discussions).
- DictatoVisit
Dictato provides local, instant voice-to-text for Mac, allowing users to transcribe speech with no cloud, account, or internet needed, supporting multiple engines and optional on-device proofreading and translation.
Pricing: 7-day free trial, then $9.99 for a two-year license.
- SpokeVisit
Spoke is a macOS app for private, on-device voice-to-text that transcribes voice into any text field using a local speech model, with an option to connect an AI provider for additional processing.
Pricing: One-time purchase (specific price not listed on Product Hunt, but website implies it).
- DictlyVisit
Dictly is a fully on-device dictation app for macOS and iOS, offering fast, private transcription with persistent custom dictionaries and optional on-device AI cleanup.
Pricing: Not explicitly stated, but mentioned as "private, on-device dictation — fast, styled and 100% offline" which implies a one-time purchase or subscription model common for indie apps.
- Whisper Snapper for MacVisit
Whisper Snapper offers local transcription for Mac using AI models, with features like speaker labels, timestamps, and various export options, or cloud transcription with user's own API keys.
Pricing: Try free or one-time lifetime $9.99 Pro.
- Session PilotVisit
Session Pilot is an offline speech transcription application for macOS designed for privacy and reliability, converting live or recorded audio into text entirely on-device.
Pricing: Not explicitly stated, but emphasis on privacy and offline use implies a one-time purchase model.
What they charge
Recent news
- Deepgram Raises $130M Series C at $1.3B Valuation to Power the Voice AI Economy
Deepgram, January 13 2026
- Deepgram Raises $130 Million Series C
The SaaS News, January 20 2026
- Deepgram Raises $130 Million Series C to Scale Real Time Voice AI Infrastructure
FNEX, January 21 2026
- Deepgram Raises $130M Series C Round at $1.3B Valuation | Built In San Francisco
Built In San Francisco, January 15 2026
- AssemblyAI pricing: Is it worth it or consider Gladia? (January 2026)
Gladia, February 04 2026
Market signals
The market for local, privacy-focused speech-to-text on macOS appears to be a growing niche, driven by increasing awareness of data privacy and the advancements in on-device AI models like Whisper. While large cloud-based solutions like Deepgram and AssemblyAI are securing significant funding rounds (Deepgram $130M Series C in Jan 2026, AssemblyAI $50M Series C in Dec 2023), indicating a robust overall AI speech recognition market, there's a clear emergence of smaller, independent developers focusing specifically on local processing for privacy and low latency on macOS.
What frustrates people
- Otter.ai users complain about reduced monthly transcription minutes on the Pro plan without a corresponding price drop, and limits on importing audio/video files unless upgrading to more expensive business plans.
- Users express frustration with Otter.ai's pricing model feeling outdated and pushing for upsells, especially with features locked behind higher-tier plans.
- Some users find that while cloud-based solutions like AssemblyAI have a low base cost for simple transcription, the cost quickly increases with advanced features due to add-on pricing.
Semble – Code search for agents that uses 98% fewer tokens than grep
Hey HN! We (Stephan and Thomas) recently open-sourced Semble. We kept running into the same problem while using Claude Code on large codebases: when the agent can't find something directly, it falls back to grep, reading full files or launching subagents. This uses a lot of tokens, and often still misses the relevant code. There are existing tools for this, but they were either too slow to index on demand, needed API keys, or had poor retrieval quality.Semble is our solution for this. It combines static Model2Vec embeddings (using our latest static model: potion-code-16M) with BM25, fused via RRF and reranked with code-aware signals. Everything runs on CPU since there's no transformers involved. On our benchmark of ~1250 query/document pairs across 63 repos and 19 languages, it uses 98% fewer tokens than grep+read and reaches 99% of the retrieval quality of a 137M-parameter code-trained transformer, while being ~200x faster.Main features:- Token-efficient: 98% fewer tokens than grep+read- Fast: ~250ms to index a typical repo on our benchmark, ~1.5ms per query on CPU (very large repos may take longer)- Accurate: 0.854 NDCG@10, 99% of the best transformer setup we tested- MCP server: drop-in for Claude Code, Cursor, Codex, OpenCode- Zero config: no API keys, no GPU, no external servicesInstall in Claude Code with: claude mcp add semble -s user -- uvx --from "semble[mcp]" sembleOr check our README for other installation instructions, benchmarks, and methodology:Semble: https://github.com/MinishLab/sembleBenchmarks: https://github.com/MinishLab/semble/tree/main/benchmarksModel: https://huggingface.co/minishlab/potion-code-16MLet us know if you have any feedback or questions!
AI
Postgres extension for BM25 relevance-ranked full-text search
Last summer we faced a conundrum at my company, Tiger Data, a Postgres cloud vendor whose main business is in timeseries data. We were trying to grow our business towards emerging AI-centric workloads and wanted to provide a state-of-the-art hybrid search stack in Postgres. We'd already built pgvectorscale in house with the goal of scaling semantic search beyond pgvector's main memory limitations. We just needed a scalable ranked keyword search solution too.The problem: core Postgres doesn't provide this; the leading Postgres BM25 extension, ParadeDB, is guarded behind AGPL; developing our own extension appeared daunting. We'd need a small team of sharp engineers and 6-12 months, I figured. And we'd probably still fall short of the performance of a mature system like Parade/Tantivy.Or would we? I'd be experimenting long enough with AI-boosted development at that point to realize that with the latest tools (Claude Code + Opus) and an experienced hand (I've been working in database systems internals for 25 years now), the old time estimates pretty much go out the window.I told our CTO I thought I could solo the project in one quarter. This raised some eyebrows.It did take a little more time than that (two quarters), and we got some real help from the community (amazing!) after open-sourcing the pre-release. But I'm thrilled/exhausted today to share that pg_textsearch v1.0 is freely available via open source (Postgres license), on Tiger Data cloud, and hopefully soon, a hyperscalar near you:https://github.com/timescale/pg_textsearchIn the blog post accompanying the release, I overview the architecture and present benchmark results using MS-MARCO. To my surprise, we were not only able to meet Parade/Tantivy's query performance, but exceed it substantially, measuring a 4.7x advantage on query throughput at scale:https://www.tigerdata.com/blog/pg-textsearch-bm25-fu
AI
Consus (YC F24) – LLM gateway for companies handling sensitive data
Hey Hacker News! I'm Eric, the founder of Consus. We route all AI requests to ITAR restricted and controlled unclassified information (CUI) compliant infrastructure. Consus itself satisfies ITAR restrictions, and we recently completed our FedRAMP equivalency audit against the 410 FedRAMP High controls. FedRAMP High is the U.S. government’s strictest security baseline for cloud services handling sensitive unclassified data. A breach of this data could cause severe or catastrophic harm. Achieving FedRAMP equivalency, or authorization, is required before processing CUI information.Demo: https://youtu.be/6MYxJe6P4oYAerospace and defense is experiencing rapid growth. Unfortunately, most companies are struggling to use AI because of necessary data protections & compliance risk. These companies can move even faster if they could use the AI tools that most other industries take for granted.I spent most of my career prior to Consus in aerospace & defense. I launched national security payloads at SpaceX, and was an early employee at Castelion. I’ve watched first hand how compliance red-tape makes it difficult for companies to adopt new technologies and trust vendors that touch their sensitive data. For example, I worked with a satellite manufacturer that doesn’t allow any AI tools, even for non-sensitive work.Even as AI has become extremely useful over the past 9-12 months, I was surprised by how many aerospace & defense companies refused to touch it. They could see how it could be helpful, but they were scared about model providers training on their data, violating an ITAR rule, or risk losing new business because of a failed cyber assessment. These are all rational fears.There were a lot of ways to solve some/all of these problems, but we focused on building a compliant AI inference layer that can "plug" into any AI tool or workflow. Just as OpenRouter integrates with most tools, Consus can connect to tools like ChatGPT or Clau
AI
Questions about this idea?
FlyBot reads the scoring and gives you a second opinion on “Ghost Pepper – Local hold-to-talk speech-to-text for macOS”.