Synopsule

Product Introduction

1. Definition: Synopsule is an on-device AI transcription and conversation analysis application for macOS and iOS. It is a native software product that records, transcribes, diarizes (labels speakers), and summarizes audio conversations entirely locally on Apple hardware using Apple's Core Audio framework and OpenAI's Whisper model.
2. Core Value Proposition: It provides a secure, private alternative to cloud-based meeting notetakers by ensuring all audio processing, transcription, and speaker identification happens locally on the user's Mac or iPhone. The core promise is professional-grade, searchable transcripts with speaker labels, delivered without requiring an account, uploading audio, or incurring a recurring subscription fee.

Main Features

1. On-Device Whisper Transcription: Synopsule bundles the Whisper AI model, enabling fully offline speech-to-text transcription that runs directly on the user's hardware. It works from the first launch without a download, processing audio captured from the microphone or system audio on macOS. Users can download larger Whisper models for enhanced accuracy, with support for Apple silicon optimization for near real-time performance.
2. Automatic Speaker Diarization & Voiceprint Recognition: The app uses a bundled on-device neural voice model to separate and label speakers in a recording (diarization). It creates local voiceprints to recognize returning speakers across multiple recordings. Users can correct or rename a speaker, and the app learns the correction, applying it universally to all relevant transcripts and future identifications.
3. Dual-Channel Capture (Mac) & Local File Management: On Mac, Synopsule utilizes Apple Core Audio taps to independently record the user's microphone and system audio (e.g., both sides of a Zoom, Meet, or Teams call) without screen recording, virtual drivers, or a meeting bot. All audio is saved to a temporary file, processed locally, and then the raw capture is deleted, leaving only a playback copy. The full audio is retained on-device for replay synced to the transcript.
4. Opt-In AI Summaries & Open Export Formats: Summaries are not automatic; users must press a button to generate them, ensuring transcript text is the only data that could ever leave the device. Summaries can be templated (e.g., standup, interview, clinical SOAP) and powered by a user's own API key (OpenAI, Anthropic, Gemini), a fully local model via Ollama or Apple Intelligence, or the optional Synopsule Pro service. Transcripts and summaries export cleanly to Markdown, Word, PDF, HTML, SRT/VTT, and integrate directly with Obsidian and Apple Notes.
5. Cross-Platform Unity & Privacy Architecture: The same codebase runs on Mac and iPhone, offering a consistent experience. The privacy model is architectural: raw audio is deleted post-transcription, summaries are opt-in, and no account is required for core functionality. All processing is local by default, with the only potential external data transfer being transcript text for summaries, which is explicitly user-controlled.

Problems Solved

1. Privacy and Data Security Concerns: Addresses the significant privacy risk associated with cloud-based meeting notetakers (like Otter, Granola, Fathom) that upload and often retain user audio recordings on remote servers. Synopsule eliminates this by keeping all audio and processing strictly on the user's device.
2. Cost and Vendor Lock-in of SaaS Models: Solves the problem of expensive, perpetual per-seat monthly subscriptions ($10-$30+/mo) for note-taking apps. Synopsule offers a one-time purchase ($1.99) for both Mac and iPhone, with optional AI summary features that users can power themselves or via an affordable add-on subscription.
3. Inconvenience of Cloud-Dependent Tools: Eliminates the need for meeting bots (like those from Otter or Fireflies) to join calls, which can be distracting or blocked by security policies. It also solves the problem of unreliable transcription in offline or low-connectivity environments by functioning fully from the first launch without internet.
4. Fragmented Notes and Lack of Searchability: Provides a unified, searchable library for all voice conversations, from Zoom calls to in-person interviews. The full-text search with speaker filters solves the pain point of finding specific moments or statements buried in hours of audio or scattered note documents.

Unique Advantages

1. Differentiation from Cloud Notetakers: Unlike competitors that are cloud-first and require accounts, bots, and ongoing uploads, Synopsule is privacy-by-architecture. It processes locally, retains full audio for replay (competitors often discard it), exports to open Markdown formats (not locked-in apps), and uses a one-time purchase model instead of a subscription.
2. Key Innovation: The core innovation is the deep integration of on-device AI (Whisper for transcription, a neural model for diarization) with Apple's system frameworks (Core Audio taps, Apple Intelligence support) to create a professional studio-grade recording and analysis tool that never compromises user privacy. The built-in MCP server for direct integration with local AI agents (Claude, Codex, ChatGPT) is another significant technical differentiator, allowing for seamless, private AI interaction with transcript libraries.

Frequently Asked Questions (FAQ)

1. How does Synopsule compare to Otter.ai or Fathom for privacy? Synopsule processes all audio and transcription completely on your Mac or iPhone, meaning your audio recordings are never uploaded to any server. In contrast, tools like Otter and Fathom require you to upload audio to their cloud for processing. With Synopsule, you maintain full, local control of your sensitive conversation data.
2. Can I use Synopsule for both Zoom calls on my Mac and in-person interviews on my iPhone? Yes. The Mac app records both sides of a Zoom/Meet/Teams call via system audio taps without a bot. The iPhone app is optimized for recording room audio via the microphone, perfect for interviews, lectures, or meetings. Both share the same on-device transcription, speaker labeling, and export features under one purchase.
3. Do I need an internet connection to transcribe meetings with Synopsule? No. Synopsule works completely offline. A Whisper transcription model is bundled with the app, so you can start recording and transcribing immediately upon installation without an internet connection, making it reliable for travel or restricted environments.
4. What happens if I never want to use AI summaries? Synopsule is designed so that if you never use the summary function, no transcript text ever leaves your device. AI summaries are strictly opt-in. You can record, transcribe, label speakers, search, play back audio, and export clean transcripts without ever connecting to any external service.