Transcribe

Welcome to the Transcribe AI service! Transcribe is designed to easily convert spoken word via audio files into text.

Getting Started

Live Site: https://ai.ccv.brown.edu/transcribe

Logging in

You will be utilizing your Brown e-mail address to log in to the ChatCCV system, same as if you were logging into Gmail or other Google services.

  • Click Sign in with Google.

  • Log in with your Brown e-mail address.

  • Select Continue to give CCV AI Services (ChatCCV) appropriate permissions to use the site.

Brown e-mail accounts are the only accounts able to access the Transcribe service.

The Transcribe Page

This page has three sections:

  • Usage Statistics, for generalized information on CO2 emissions from utilizing the service, and other service-related facts.

  • FAQ, for frequently asked questions on the service.

  • All Jobs, for access to Create Job and viewing status or results of Jobs you have requested to run.

Creating a Job

Click Create Job under All Jobs to open the Create Job slide-out.

  • Step 1: Start uploading files by clicking on the file upload zone or dragging and dropping files to it. You can delete the files that you have uploaded by mistake. We have the following limits of the files that you upload:

    • It must be common audio and video files

    • The service is designed for conversational content. The models might not perform well on non-conversational content.

    • Size limit: 1GB per file

    • Duration limit: 6 hours total

As of June 23, 2025, Transcribe now supports most audio and video files. However, some files, especially video files, are of an encoding that's not supported by the browser and will be rejected. If that happens, please extract the audio from your video file and/or convert the audio to a popular format such as wav or mp3 for compatibility.

  • Step 2: Select a model. You currently have 3 models to choose from:

    • Google's Gemini 2.5 model, which is Google's flagship AI model that is capable of handling audio transcription tasks. Though currently experimental, it can produce surprisingly good results.

    • OpenAI's Whisper model, though created by OpenAI, is run on a Brown-maintained service, not on OpenAI servers.

    • Microsoft Azure's AI Speech: Speech to Text model is run on Microsoft servers under Brown's Use Agreement with them.

We have plans to review and add new models throughout the lifetime of this service.

  • Step 3: Choose a name for the job. If not supplied, a name is automatically determined from the names of the files.

  • Step 4: Locale selects a language or dialect. Different models may have different options.

  • Step 5: Click on Start Job to send the files for transcription - it's semi-transparent when not all pre-requisites are complete.

Record retention policy

Brown must store your data on your behalf to do transcription services for you. For more information on data retention, see Transcribe's Privacy Statement.

Last updated

Was this helpful?