Skip to main content
AI Processing polishes transcribed text by fixing grammar, adding punctuation, and correcting filler words before inserting it into your app.
When enabled, transcribed text (not audio) is sent to your chosen AI provider. Avoid using with sensitive content.

Quick Setup

Enable AI Processing in Settings → AI Processing, select a provider, and enter your API key (OpenRouter only). For local/custom endpoints, provide a base URL and model name; the API key can be left blank.

Supported Providers

AI Modes

ModeDescription
Transcription EnhancerFixes grammar and punctuation while preserving your natural voice
AI AssistantRewrites text with improved tone and structure

Context-Aware Enhancement

Pipit’s AI detects what you’re writing and applies specialized formatting:
ContextWhat It Does
EmailFormal greetings, proper closings, professional tone
Code/TerminalBacktick code formatting, converts “cd into src” to cd src/
Git Commands”git commit dash m fix” becomes git commit -m "fix"
Chat/MessagesCasual tone preservation, emoji support
Social MediaMinimal changes, preserves your exact voice
SpreadsheetData-optimized formatting

Creative Use Cases

Privacy-Sensitive Dictation

Use local AI (Ollama) for medical notes, legal documents, or any content that shouldn’t leave your machine.

Offline Enhancement

Traveling or on a plane? Local AI keeps working without internet.

Custom Vocabulary

Configure local models with custom system prompts for domain-specific terminology.

Fallback Behavior

If AI processing fails, raw transcription is inserted with a notification. Your dictation is never lost.

Cost

Cloud: Uses your own API key through OpenRouter (google/gemini-2.5-flash-lite is affordable). Local: One-time Ollama setup, then free inference forever.