Product Introduction
Definition: WizardNote is an offline-first voice note transcription and AI-powered summarization mobile application designed for iOS 18+ and Android 10+. Categorized under productivity software, voice-to-text transcription utilities, and markdown note-taking assistants, the app executes local speech-to-text processing directly on the user's mobile hardware using on-device models such as Whisper.
Core Value Proposition: WizardNote exists to provide secure, private voice dictation and meeting transcription without transmitting raw audio files to external cloud servers. By leveraging local processing, built-in vocabulary learning, and local privacy filters that strip sensitive personally identifiable information (PII) like emails, phone numbers, and street addresses prior to any optional AI interaction, WizardNote guarantees absolute data privacy while converting spoken thoughts into structured markdown files, actionable task checklists, and Obsidian-compatible meeting logs.
Main Features
On-Device Offline Whisper Transcription: The speech recognition engine runs locally on the mobile processor using Whisper architecture. This allows WizardNote to operate entirely offline without internet connectivity, supporting multi-language processing (English, Spanish, German, French, and Czech) with selectable model sizes ranging from basic to high-accuracy variants (requiring between 190 MB and 1.6 GB of local storage).
Local PII Stripping and Privacy Shielding: Before any textual transcript is routed to an external artificial intelligence provider for summarization, WizardNote scans and strips sensitive data elements including email addresses, phone numbers, physical addresses, and (in Pro mode) specific names and corporate identifiers. Raw audio recordings never leave the user's mobile device under any circumstance.
Flexible AI Summarization and Modular Backends: Users retain complete control over how their transcripts are synthesized into summaries and tasks. Summaries can be generated using local processing via Apple Intelligence or Gemini Nano, custom API keys (OpenAI, Anthropic), a self-hosted server backend, free manual copy-pasting into ChatGPT/Claude/Gemini, or via an optional privacy-compliant EU cloud subscription.
Obsidian-Ready Markdown Export and Task Automation: Every recorded session compiles into a structured plain-text markdown file containing frontmatter tags, identified speakers, chronological transcripts, and extracted action items. Pro users can instantly sync these tasks directly to native iOS Reminders, system Calendars, Apple Notes, or their local Obsidian vault via deep system shortcuts.
Universal Capture via Hardware Shortcuts: Recording can be initiated instantly with a single tap from anywhere on the device. Supported trigger methods include the iPhone Action Button, home screen widgets, Siri, Android quick settings tiles, and the ability to import pre-recorded audio files, phone call exports, or Zoom/Teams meeting recordings via the system share sheet.
Problems Solved
Pain Point: Privacy vulnerabilities and data leakage associated with mainstream cloud-based voice dictation and AI transcription apps that upload proprietary business or personal audio recordings to third-party servers. WizardNote solves this by eliminating mandatory user accounts, telemetry, and audio transmission entirely.
Target Audience: Privacy-conscious professionals, software developers, journalists, researchers, project managers, and Obsidian knowledge-management enthusiasts who require secure meeting transcription without sacrificing local data ownership.
Use Cases: Rapidly capturing voice memos during commutes without internet access; recording confidential client meetings with automatic PII redaction; importing external call recordings from Microsoft Teams or Zoom for structured markdown archiving; and automatically parsing spoken task items directly into project management workflows or personal knowledge bases.
Unique Advantages
Strengths & Limitations (Pros & Cons):
- Pros: Zero audio data leaves the phone; works completely offline; supports one-time Pro purchase model (no mandatory subscription); integrates natively with Obsidian and Apple/Android task ecosystems; customizable AI backends allow users to utilize their own API keys or local servers.
- Cons: High-accuracy transcription models require substantial local storage space (up to 1.6 GB); advanced speaker identification, custom templates, and direct task-sync require the Pro tier purchase; automated cloud summaries require a monthly/yearly subscription if users do not want to configure their own API keys.
Key Alternatives & Differentiation: Compared to cloud-dependent transcription tools like Otter.ai or standard cloud dictation suites, WizardNote differentiates itself through its strict zero-audio-upload policy and local Whisper processing. Unlike traditional voice recorders that output raw, unsearchable audio files, WizardNote structures unstructured speech into actionable, tagged markdown notes designed specifically for local-first knowledge management apps like Obsidian, offering a hybrid model of local security and modular AI flexibility.
Frequently Asked Questions (FAQ)
Does WizardNote send my voice recordings to the cloud? No. All speech-to-text transcription runs locally on your iPhone or Android device using on-device models. Raw audio recordings never leave your phone under any circumstances.
How does WizardNote handle sensitive information like phone numbers and emails? Before any text is sent to an optional AI summarization service, WizardNote automatically scans and strips sensitive data such as email addresses, phone numbers, and physical addresses. Pro users can also redact specific names and corporate identifiers.
Do I need an internet connection to record and transcribe voice notes? No. Once you download the initial transcription model (ranging from 190 MB to 1.6 GB depending on accuracy and platform), transcription functions entirely offline without requiring an active internet connection.
Can I use WizardNote without paying a monthly subscription? Yes. The core recording, offline transcription, and basic markdown export features are entirely free. You can use free manual summaries by pasting transcripts into ChatGPT, or purchase the Pro version with a single one-time payment to unlock advanced task integrations, custom templates, and direct AI API key support.
Which file formats and productivity apps are supported for exports? WizardNote exports recordings as standard plain-text Markdown files (
.md) containing YAML frontmatter tags, speaker lists, structured summaries, and checkbox task lists. These files integrate seamlessly with Obsidian, Apple Notes, Apple Reminders, system calendars, and file management apps via system share sheets and shortcuts.
