Field Theory

Product Introduction

1. Definition: Field Theory is a macOS utility application designed for AI context management, specifically targeting voice transcription, screenshot annotation, and portable command execution. It functions as a local productivity layer for developers and engineers interacting with AI models.
2. Core Value Proposition: Field Theory eliminates the tedious manual context stacking required when using AI tools like ChatGPT, Claude, or Cursor. It enables users to capture, process, and inject context (voice, screenshots, commands) into any input field via the clipboard, doubling context efficiency while halving workflow steps.

Main Features

1. Portable Commands:
   • How it works: Users define custom commands in local Markdown files within any folder. Field Theory indexes these commands, making them globally accessible via the system-wide hotkey command + shift + k. Selecting a command instantly pastes its content into the active input field (e.g., Claude, Cursor, browser-based AI).
   • Technology: Local file system indexing, macOS global keyboard shortcuts, clipboard manipulation.
2. Priority Mic:
   • How it works: Field Theory locks the user's preferred microphone input source at the system level. It overrides macOS's default behavior of switching inputs when Bluetooth devices connect, ensuring consistent audio capture for transcription.
   • Technology: macOS Core Audio API hooks for persistent input device selection.
3. Local Voice Commands & Transcription:
   • How it works: Voice input is transcribed locally on-device using Apple's Neural Engine. During transcription, users can verbally invoke Portable Commands (e.g., "Run the deslop command"). The spoken command name triggers Field Theory to insert the corresponding Markdown content seamlessly into the transcript.
   • Technology: Apple's on-device Speech Recognition framework (macOS 14+), custom command-trigger parsing.
4. Draw on Screenshots (Local Annotation):
   • How it works: Captured screenshots open instantly in Field Theory's built-in editor. Users can add arrows, circles, freehand drawings, or text annotations. The annotated image is copied directly to the clipboard with a single click, ready for pasting into any AI input field.
   • Technology: Native macOS image capture, custom raster graphics editor, clipboard integration.
5. Auto-Improve Text (Cloud):
   • How it works: Transcribed voice input or pasted text can be processed to remove filler words (ums, ahs), tighten phrasing, and automatically label attachments (e.g., converting "I added a screenshot" to "[Figure 1]"). This transforms spoken prompts into concise, written-quality input.
   • Technology: Cloud-based NLP processing (requires internet).

Problems Solved

1. Pain Point: Inefficient context management for AI interactions, requiring excessive manual copying/pasting of logs, docs, code snippets, and screenshots between tools.
2. Target Audience: Software Developers, DevOps Engineers, Data Scientists, Technical Founders, and System Architects who frequently use AI coding assistants (Cursor, GitHub Copilot) or LLMs (Claude, ChatGPT) for problem-solving and need to provide rich context.
3. Use Cases:
   • Debugging: Quickly pasting annotated screenshots of errors + relevant logs into an AI chat.
   • Code Generation: Verbally describing a function requirement while invoking a "unit test template" command mid-sentence.
   • Documentation: Transcribing voice notes about a system design and having them auto-formatted with clear attachment references.
   • Meetings: Reliably capturing voice notes locally without Bluetooth interference.

Unique Advantages

1. Differentiation: Unlike browser extensions or API-dependent tools, Field Theory is a native macOS app using the clipboard as its integration point, making it input-field agnostic and compatible with any application (ChatGPT, Cursor, Gemini, Perplexity, IDE, notes app). Competitors typically lock users into specific platforms or require complex setups.
2. Key Innovation: The combination of fully local, offline-capable features (Priority Mic, Local Transcription/Commands, Screenshot Drawing) with seamless clipboard-based context injection creates a unique zero-friction workflow. The ability to verbally trigger complex commands during live transcription is a significant productivity breakthrough for technical users.

Frequently Asked Questions (FAQ)

1. Does Field Theory require an internet connection?
   Core features (Portable Commands, Priority Mic, Local Voice Transcription/Commands, Screenshot Drawing) work entirely offline. Only the "Auto-Improve Text" feature requires an internet connection for cloud-based NLP processing.
2. What are the system requirements for Field Theory?
   Field Theory requires macOS 14 Sonoma or later and runs exclusively on Apple Silicon Macs (M1, M2, M3 chips or later). This ensures optimal performance for its local AI processing (transcription) and system integration.
3. How do Portable Commands work with different AI tools?
   Portable Commands are folder-based Markdown files stored locally. Field Theory makes them accessible via a global hotkey (command + shift + k). Selecting a command copies its content to your clipboard. You then simply paste (command + v) it into any input field (Cursor, Claude, ChatGPT, email, etc.) – no tool-specific integrations needed.
4. Is there a free version of Field Theory?
   Yes, Field Theory offers a feature-rich Free (Basic) tier. It includes unlimited local transcription, screenshots, drawings, manual stacking, single image stacking, The Librarian, and priority mic. The Free tier limits Portable Command uses (100/month), Priority Mic minutes (500/month), Auto-stacks (50/month), and Auto-improve words (5,000/month). The Pro tier ($14/month) removes these limits and increases Auto-improve to 50,000 words/month.
5. Can I use Field Theory with my Bluetooth headphones?
   Absolutely. The Priority Mic feature is specifically designed for this. Set your preferred mic (e.g., headset mic) in Field Theory once. It will prevent macOS from automatically switching the input source when your headphones connect/disconnect, ensuring your chosen mic stays active for reliable transcription.