The Choicer Voicer
Browser voice impression challenge requiring a microphone.
Browser-based speech-to-text for meetings, podcasts, and interviews.
Gemini 3.5 Transcribe is an advanced, browser-based speech-to-text solution designed to transform audio recordings—such as meetings, podcasts, interviews, and voice notes—into precise, searchable, and editable transcripts. Powered by Google's Gemini 3.5 model yet functioning independently, this tool expertly handles complex audio environments including multi-speaker conversations, noisy backgrounds, and code-switching between over 85 languages. Users can upload files up to 6 hours or 2 GB in size or record audio directly within the browser, facilitating a hassle-free transcription process.
Offering two distinct transcription modes, Smart and Verbatim, Gemini 3.5 Transcribe adapts to different user needs: Smart mode delivers polished, cleaned-up text with filler words removed, while Verbatim mode preserves the exact spoken content with detailed timestamps and speaker labels. Its robust speaker diarization supports up to 8 participants, and the custom vocabulary option allows users to tailor the transcription accuracy by including specialized terms and names. With versatile export formats like TXT, SRT, VTT, and JSON, Gemini 3.5 Transcribe fits seamlessly into diverse workflows, from captioning and publishing to archiving and automation.
Gemini 3.5 Transcribe offers a straightforward minute-based pricing model with two subscription tiers. The Starter Plan is priced at $9.90 per month, providing 300 transcription minutes and access to the Gemini 3.5 model with standard email support, ideal for light or occasional users. The Pro Plan costs $24.90 per month and includes 1,200 transcription minutes, plus additional features like speaker labels, word-level timestamps, and priority access to the transcription queue, tailored for content creators, reporters, and growing teams. Additional minutes can be purchased as needed, making the pricing flexible and scalable according to transcription volume.
Pros
Cons
It supports automatic detection across more than 85 languages and can handle language switching within a single recording. Users can also provide a language hint to improve transcription accuracy if the language is known in advance.
Gemini 3.5 Transcribe can process audio files up to 6 hours in length or 2 GB in size. Speaker labels are available for recordings up to 30 minutes; longer files are automatically split and transcribed seamlessly within the user's account.
Smart mode produces polished text by removing filler words, resolving self-corrections, and formatting content for easy reading. Verbatim mode keeps the transcription exactly as spoken, preserving word-level timestamps and speaker labels, which is essential when exact quotes or detailed attribution are required.
Yes, it supports speaker diarization for up to eight speakers. However, labeling conversations with three or more speakers is still considered experimental and may have reduced accuracy.
You can export your transcriptions as plain text (TXT), subtitles in SRT or VTT formats, and structured JSON files. Subtitle timing features require verbatim mode with word timestamps enabled.
Audio files are securely sent through the Vercel AI Gateway to Google's Gemini 3.5 model for transcription. The resulting transcript tasks are stored in your account, and storage policies can be configured according to deployment requirements to ensure data privacy and compliance.
Need to organize your AI tool files?
Managing files from Gemini 3.5 Transcribe and other tools? The Drive AI automatically organizes, tags, and retrieves all your files with AI.
Try The Drive AI freeBrowser voice impression challenge requiring a microphone.
Generates royalty-free songs from text prompts, lyrics, or audio.
Free online tool to separate vocals from audio recordings.