# Dicty.io — Free Voice-to-Text & Architectural Knowledge Engine > Dicty is a free, native cross-platform desktop voice-to-text application and architectural documentation engine for macOS and Windows. It combines zero-latency on-device speech transcription via Whisper.cpp (Metal GPU accelerated on Apple Silicon, AVX2 on Windows) with optional cloud/local LLM intelligence to transform spoken thoughts into perfectly formatted code, pull request descriptions, Architecture Decision Records (ADRs), and synchronized team documentation. Dicty operates as an open, private alternative to subscription-locked dictation tools like Superwhisper, MacWhisper, and Wispr Flow. Users can dictate completely offline with local Whisper models, bring their own API keys (BYOK) for frontier models (Claude, OpenAI, Gemini, DeepSeek, Qwen), or connect to local Ollama endpoints with zero recurring subscription fees. ## Core Capabilities - System-Wide Voice-to-Text HUD: Floating menu bar overlay (Option+Space on macOS, Ctrl+Space on Windows) with real-time streaming transcription and instant auto-paste into Cursor, VS Code, Slack, Obsidian, Notion, and browsers. - Local & Cloud Speech Engines: Multi-model support ranging from Whisper Tiny and Base to Large-v3 and Large-v3 Turbo, fully hardware-accelerated. - Text-to-Voice (Read Aloud): Offline text-to-speech synthesis powered by local Kokoro TTS and Silero engines, allowing instant audio readouts of selected text. - Audio File Transcription: Drag-and-drop lecture, interview, and voice memo transcription with automated AI structuring, key takeaways, and action items. - Architectural Grounding: Voice-driven generation of numbered Architecture Decision Records (ADRs), clean PR descriptions, Git commit messages, and structured technical specs. - Knowledge Vault Sync: Real-time documentation synchronization into Obsidian vaults, Notion databases, Confluence, and GitHub wikis. - Multilingual Voice Translation: Real-time speech translation across 90+ spoken languages with on-the-fly target language insertion. - Custom Prompting & Profiles: Context-aware prompt pipelines (Markdown, Jira, Bug Report, Meeting Minutes, Executive Summary) with domain-specific vocabulary and phonetic dictionary rules. ## Documentation - [Getting Started](https://dicty.io/docs/getting-started): Installation guide for macOS (Homebrew, DMG) and Windows, system requirements, microphone and accessibility permissions, and first dictation walkthrough. - [Speech Engines & Models](https://dicty.io/docs/models-and-speech-engine): Comprehensive comparison of local Whisper models (Tiny to Large-v3 Turbo), Metal GPU acceleration, memory requirements, and custom cloud endpoints. - [AI Customization & Prompts](https://dicty.io/docs/ai-customization-and-prompts): System prompt engineering, custom instructions, tone control, formatting rules, and dynamic variables. - [File Transcription](https://dicty.io/docs/file-transcription): How to transcribe audio and video files (MP3, WAV, M4A, MP4) with speaker diarization and AI summarization for student lectures and meetings. - [Modes & Profiles](https://dicty.io/docs/modes-and-profiles): Switching between developer, documentation, casual, and custom prompt presets. - [Knowledge Base & Vault Sync](https://dicty.io/docs/knowledge-base-and-sync): Bi-directional sync with local Markdown folders, Obsidian vaults, and cloud workspaces. - [Project Context & GitHub](https://dicty.io/docs/project-context-and-github): Generating Git commits and PR descriptions directly from spoken context and repository diffs. - [Voice Translation & Multilingual](https://dicty.io/docs/voice-translation-and-multilingual): Dictating in one language and typing out in another in real-time. - [Menubar & Overlay HUD](https://dicty.io/docs/menubar-and-overlay): Floating overlay configuration, hotkeys, status animations, and menu bar controls. - [Settings & Preferences](https://dicty.io/docs/settings-and-preferences): Audio input devices, sound feedback, telemetry preferences, and updates. - [Subscription Tiers & Limits](https://dicty.io/docs/subscription-tiers-and-limits): Free tier vs. Pro vs. Early Supporter lifetime license, BYOK model explanation. ## Competitor Comparisons - [Dicty vs. Superwhisper](https://dicty.io/compare/dicty-vs-superwhisper): Comparison of features, pricing, offline capabilities, model freedom, and developer workflows. - [Dicty vs. MacWhisper](https://dicty.io/compare/dicty-vs-macwhisper): Comparison of system-wide dictation, auto-paste, LLM post-processing, and cross-platform support. - [Dicty vs. Wispr Flow](https://dicty.io/compare/dicty-vs-wispr-flow): Local privacy vs. mandatory cloud routing, subscription costs, and offline usability. - [Dicty vs. Apple Dictation](https://dicty.io/compare/dicty-vs-apple-dictation): Punctuation accuracy, programming vocabulary support, and LLM formatting capabilities. - [Dicty vs. Otter.ai](https://dicty.io/compare/dicty-vs-otter-ai): Local audio privacy, developer-centric documentation workflows, and pricing differences. ## Free Developer Tools - [Voice-to-Markdown](https://dicty.io/tools/voice-to-markdown): In-browser web tool to convert voice speech directly into clean, formatted GitHub-flavored Markdown. - [ADR Template Engine](https://dicty.io/tools/adr-template-engine): Web tool to generate standard Architecture Decision Records from brief problem descriptions. - [Git Diff Summarizer](https://dicty.io/tools/git-diff-summarizer): Paste a git diff or speak your changes to generate concise, conventional commit messages and PR bodies. - [Web Voice Tester](https://dicty.io/tools/web-voice-tester): Test microphone input and live speech-to-text directly in the browser before installing the native desktop app. ## Use Cases - [Architecture Decision Records (ADRs)](https://dicty.io/use-cases/adrs): Capture architectural reasoning and tradeoffs out loud and generate linkable, numbered ADRs. - [PR & Commit Descriptions](https://dicty.io/use-cases/pr-commit-docs): Speak for 30 seconds to produce comprehensive, professional pull request descriptions. - [Developer Onboarding](https://dicty.io/use-cases/dev-onboarding): Walk through codebases and architecture verbally to create living onboarding guides. - [Student Lecture Notes](https://dicty.io/use-cases/student-lecture-notes): Convert classroom audio recordings into formatted study summaries with AI takeaways. - [Speech-to-Text for Coding](https://dicty.io/use-cases/speech-to-text-for-coding): Dictate code, comments, docstrings, and terminal commands with programming identifier formatting. - [Obsidian Voice Notes](https://dicty.io/use-cases/obsidian-voice-notes): Seamless daily voice memo logging into Markdown frontmatter and second-brain vaults. - [Text-to-Voice (Read Aloud)](https://dicty.io/use-cases/read-aloud-text-to-voice): Listen to articles, code reviews, and documentation read aloud offline. - [Multilingual Voice Translation](https://dicty.io/use-cases/multilingual-voice-translation): Speak your native language and type directly in Spanish, German, Russian, or English. ## Installation & Downloads - [Download Dicty Desktop](https://dicty.io/download): Download native installers for macOS (Apple Silicon & Intel) and Windows x64. - [GitHub Releases](https://github.com/dicty-io/dicty-releases/releases/latest): Direct binary assets, SHA-256 checksums, and changelogs. - [Homebrew Tap](https://github.com/dicty-io/homebrew-dicty): `brew install --cask dicty-io/dicty/dicty`. ## Community & Support - [Discord Community](https://discord.gg/CNEkeVj3b): Official Discord server for technical support, feature requests, and developer discussions. - [Public Roadmap](https://dicty.io/roadmap): Community feature voting and release schedule. - [Contact & Support](https://dicty.io/contact): Email `hello.dicty.io@gmail.com` for developer queries and enterprise deployments. ## Optional - [Full Documentation Aggregate (llms-full.txt)](https://dicty.io/llms-full.txt): Consolidated single-document text containing complete documentation pages and specifications for deep agent ingestion. - [Site Map](https://dicty.io/sitemap.xml): Full XML sitemap of all public routes and localized URLs.