The voice recording component captures audio through the browser's microphone API, processes it through speech recognition models, and outputs transcribed text directly into the active text field. Users click the extension icon in any text box, start recording, and watch their speech appear as written text. The transcription runs on remote servers rather than locally, which means recording quality depends on network speed and server load.
Text polishing works through a separate AI model that analyzes grammar, tone, and clarity. Users highlight existing text, trigger the polishing function, and receive suggested improvements. The system doesn't just check spelling. It restructures sentences and adjusts phrasing based on context patterns learned from training data.
Both functions consume credits from a user's account balance. Recording burns one credit per thirty seconds of audio. Polishing costs one credit per roughly 150 words of text. This creates a predictable usage pattern where a three minute recording costs six credits, and polishing a 600 word draft costs four credits.
Processing speed varies by subscription tier. Standard speed routes requests through regular processing queues. Priority speed jumps ahead in line but still shares infrastructure with other priority users. Fastest lane gets dedicated processing resources with minimal wait time. The difference matters during peak usage when standard requests might queue for several seconds.
The extension doesn't integrate with specific platforms. It works universally across web applications because it hooks into the browser's text input handling rather than connecting to individual services through APIs. This means it functions in Gmail, Slack, Google Docs, content management systems, and any other web interface with text fields.
Free accounts start with 100 welcome credits plus 50 monthly credits that reset. Standard plans provide 750 credits monthly for $4, which translates to roughly twelve notes per day if each note involves both recording and polishing. Pro tier costs $8 monthly for 2,500 credits, supporting around forty notes daily. Unlimited plans run $14 monthly with uncapped credits but include a fair use policy capping usage around 100 hours per month. Annual billing cuts costs by 33 percent across all paid tiers.
The credit system creates hard limits on free and lower paid tiers. Running out of credits mid-month means either upgrading or waiting for the monthly reset. The fair use policy on unlimited plans isn't precisely defined beyond the 100 hour guideline, which could cause friction for truly heavy users who expect completely unlimited access.
Recording quality depends on microphone input and background noise since the system processes raw audio without sophisticated noise cancellation. Accents and speech patterns outside the training data distribution may produce transcription errors. Text polishing suggestions reflect the model's training biases, which means recommended changes might not match specific style guides or industry terminology.
The extension requires Chrome or Chromium-based browsers. No mobile app exists. Pro tier includes priority support, but response times and support channels aren't specified. Processing happens on external servers, so the extension requires internet connectivity and sends audio data off-device for transcription.
