Meetily is a desktop application built for professionals and enterprises who need to record, transcribe, and summarize meetings without sending data to the cloud. Designed for macOS, Windows, and Linux, it eliminates privacy risks associated with SaaS meeting tools by processing all audio and AI tasks locally using open-source models. It’s ideal for compliance-sensitive fields like legal, healthcare, and defense where data sovereignty is non-negotiable.
Built with Rust and Tauri, Meetily combines a high-performance backend with a Next.js frontend. It integrates Whisper.cpp and Parakeet for real-time speech-to-text, Ollama for local LLM summarization, and supports custom OpenAI-compatible endpoints. The app runs offline, stores all data locally, and supports GPU acceleration via Metal, CUDA, and Vulkan for fast processing.
What You Get
- Local Real-Time Transcription - Uses Whisper.cpp and Parakeet models to transcribe live meetings with no internet connection required, ensuring zero data leakage.
- AI-Powered Summaries with Ollama - Generates meeting summaries using locally hosted LLMs via Ollama, with support for Claude, Groq, and OpenRouter as alternatives.
- Multi-Format Audio Import - Import MP4, M4A, WAV, MP3, FLAC, OGG, AAC, MKV, WebM, and WMA files to re-transcribe and enhance past recordings locally.
- Professional Audio Mixing - Simultaneously captures microphone and system audio with intelligent ducking and clipping prevention for clean recordings.
- Custom OpenAI Endpoint Support - Configure any OpenAI-compatible API endpoint for summarization, enabling use of private or custom models without cloud dependency.
- GPU Acceleration (Metal/CUDA/Vulkan) - Automatic hardware acceleration on macOS (Apple Silicon), Windows (NVIDIA), and Linux (AMD/Intel) for faster transcription and summarization.
- Offline-First Design - All recordings, transcripts, and summaries are stored locally on your device with no cloud storage or third-party data collection.
- GDPR & HIPAA Ready Architecture - Built with data sovereignty in mind; no telemetry, no external APIs, and full control over sensitive meeting data.
Common Use Cases
- Legal teams documenting client calls - Lawyers use Meetily to transcribe and summarize confidential client discussions without violating attorney-client privilege or data residency laws.
- Healthcare providers recording patient consultations - Clinicians capture and summarize patient interactions locally to comply with HIPAA while maintaining accurate medical records.
- Enterprise teams running product strategy meetings - Product managers use Meetily to generate action items and decisions from internal meetings without exposing proprietary strategy to cloud vendors.
- Remote consultants and freelancers - Independent professionals transcribe client calls on their own machines to maintain trust and avoid SaaS subscription costs.
Under The Hood
Architecture
- Clean separation of concerns with a modular frontend-backend structure, where React/Next.js components are decoupled from Rust-based backend services via well-defined API contracts
- Explicit domain modeling with dedicated layers for data representation, service logic, and configuration, enforced through typed DTOs and dependency injection patterns
- Bounded contexts maintained across frontend, Tauri native layer, and Python AI modules, minimizing cross-layer dependencies and promoting maintainability
- REST endpoints directly align with domain operations, ensuring intuitive and predictable API behavior
Tech Stack
- React and Next.js with TypeScript form the frontend foundation, enhanced with modern UI libraries for components and animations
- Tauri serves as the desktop runtime, integrating Rust for system-level access and performance-critical operations, managed via Cargo workspace
- Python microservices handle AI/LLM workflows with Pydantic and FastAPI, leveraging Ollama for local model inference
- Unified build system combines Rust’s Cargo with TypeScript tooling, enabling cohesive development across native and web layers without traditional backend frameworks
Code Quality
- Extensive testing coverage with unit and integration tests validating component behavior and state management
- Consistent code organization following React best practices, including reusable hooks and context providers for state
- Strong type safety through comprehensive TypeScript interfaces and type guards, reducing runtime errors
- Linting and formatting are strictly enforced, ensuring code consistency and readability
- Error handling is present but relies on generic try-catch patterns without custom error classes or structured boundaries
What Makes It Unique
- Fully offline-first desktop experience with direct access to audio and file systems, eliminating cloud dependency for core functionality
- Embedded AI summarization pipeline with configurable templates and real-time editing via BlockNote, enabling dynamic note generation during meetings
- End-to-end encrypted local storage ensures all data remains on-device by default, prioritizing privacy over SaaS models
- Integrated update mechanism with in-app patching provides a seamless native update experience without external installers
- Unified stack combining audio capture, transcription, and summarization into a single cohesive desktop application, not a web wrapper or plugin