Local processing

Private Voice Tools

Drop in files, ask questions, transcribe recordings, and create speech.

The local tools are free. Your device downloads the models and completes the AI work. Workspace files stay only in this browser. You can add private cloud backup later.

TranscribeRead aloud Assistant beta

Model and speaker settings

Not loaded
Checking this device Checking

The check does not upload any device data.

    Technical device details
    
                                                    
                                                
    Whisper engine
    This beta adds memory use. Review its speaker names and boundaries.
    Expected speakers
    These values check the beta result. They do not force the local model to find that count.
    1

    Add media

    Source and rights

    Public-domain English audio. The satire discusses poverty and cannibalism.

    This browser model version requires a language choice. The browser remembers it.
    Drop audio or video here or select a file from this device
    or
    Recording 00:00
    Not ready No media is ready. Next: choose an audio or video file.
    ?
    Waiting for a file.
    Working in this browser
    Preparing model
    Keep this page open during local processing.
    2

    Review transcript

    No transcript
    Transcript checks
      Rename, merge, or reassign speakers. The raw model result stays unchanged.
      Reassign a time range
      Text edits change TXT, copy, and share output. Timed exports use the detected words.

      Browser storage

      Local only

      Saved files live in this browser's private storage area on this device. Voiceless does not upload them.

      Local files do not sync between browsers or devices. The browser can remove them when storage is low.

      Anyone using this browser profile can open them. Clearing this site's browser data deletes them.

      Library
      0 records
      Browser storage
      Checking
      Retention
      Best effort

      Want backup and cross-device search?

      These legacy records do not join automatic workspace backup. Use Save to cloud when you want a private server copy.

      View cloud workspace

      Saved recordings

      Opening this browser's local library.

      Your files

      0 items
      Drop files here Files stay in this browser.
      /
      Opening this browser's workspace.
      Browser storage This browser only
      Back up files on every device

      Local assistant

      Not loaded

      Liquid LFM runs on your graphics processor. Its first download is about 200 MB. A capable computer can load it automatically. Phones load this beta only when you ask. Direct tools work without it.

      Loading model

      Private assistant

      One chat for your local files and voice tools
      Drop in a file or ask in normal language. I can read text and prepare local voice actions.
      Checking this device Chat loads first when this device can hold the local model safely.
      Try public examples Test this device without using a private file.
      Sources and rights

      The book is the complete public-domain text of Alice in Wonderland.

      Type a message. The local assistant loads before the first answer when needed.

      Ask the assistant

      Direct tools

      Direct tools do not need the chat model. Choose a text, audio, or video file to begin.

      Preparing speech

      Sign in to use the cloud library

      Local transcription stays free and does not require an account.

      Sign in