Playground

Assets and history

Find every speech generation, transcript and cleaned recording made in the workspace on the Assets page, filter and search them, play them again, download, rename and delete them.

The Assets page is the workspace's record of every Text to Speech generation, Speech to Text transcript and Sound Cleaning result made from the portal, by any member, with the newest first, plus your cloned voices. When you finish this page you will know how to find an item by its type, its dialect or a word it contains, play it again, download it, rename it and delete it. None of that costs credits.

What you need

  • A workspace you belong to. Every member sees the same list.
  • At least one generation, transcript or cleaned recording made from the portal. Until then the page reads No history yet and Nothing here yet. Generate speech, transcribe a file or clean a recording, and it lands here.

Steps

Step 1: Open Assets

In the sidebar under Playground, click Assets. The same page opens from the All in Assets → link on the Text to Speech and Speech to Text pages. The page is headed Everything you made, in one line, with the line Every voice you generated, every file you transcribed or cleaned, and every voice you added. Kept until you delete it. While the list loads, the page reads Loading history….

The page always opens on the All tab, including when you arrive from the Speech to Text page. Choose a tab in the next step to narrow it.

Five tabs sit above the list, each with a count: All, Text to Speech, Speech to Text, Sound Cleaning and Voice Cloning. Under them is a search box reading Search text, files and voices.

The search matches an item's name, the text that was spoken, the file name, the voice's name and its dialect, whatever the letter case. On the Voice Cloning tab it matches the voice's name. When nothing matches, the page reads No history matches your search and Try a different keyword or clear the search.

When the items on the tab carry dialects, a row of dialect chips appears under the search box, each with a count. Press a chip to keep only that dialect, and press it again to clear it. Changing the tab clears the chip.

The tabs and the search work on the items the page has loaded: the newest 100 of each kind in the workspace. If part of the history did not load, the page reads Part of your history did not load. with a Try again button.

Step 3: Read a row

The list is grouped by day, newest first, under Today, Yesterday, then the date, for example 15 Sep. Every item is one row:

Part What it shows
Icon For a generation, the voice's orb. For a transcript or a cleaned recording, an icon for its type.
Title The name you gave it, or else the text that was spoken or the transcript, or the file name of a cleaned recording. A renamed item shows what the name replaced under it.
Meta line The type, the output (WAV, TXT), and the time as HH:mm, followed by Failed or Processing when the item is not ready.
Voice For a generation, the voice's name and its dialect.
Length The duration, and the size of the file.
Buttons Play or Pause (a generation or a cleaned recording) or Open the transcript (a transcript), Download, Rename and Delete.

Step 4: Play an item

Press Play on the row. The player opens at the bottom of the page. One item plays at a time: pressing Play on another row replaces the first, and leaving the page stops playback. Press Pause to pause, Stop to stop it, or Close to close the player.

  • A generation shows a spinner on the button while the audio is fetched, then plays. Play is offered on Ready generations only.
  • A transcript has Open the transcript instead. It loads the audio that was transcribed into the player and shows the transcript above it. Press Close the transcript to close it. Play is offered when that audio was kept.

Step 5: Download the audio or the transcript

Press Download on the row.

Item File you get
Generation The audio, named after the item's name or the first 40 characters of its text, with the extension of the audio, which is .wav for a portal generation. See Formats, quality and download.
Transcript The transcript in the format you pick, TXT, SRT or JSON, named after the item's name, or else the uploaded file's name without its extension, with .txt, .srt or .json. SRT cues and JSON speakers carry the speaker labels, in the language of the page.
Cleaned recording The audio, named after the item's name or the cleaned file's name, as .wav.

The audio of a transcript can be played here but not saved from here.

Download is greyed out on a Failed or Processing row. If the file could not be fetched, the row reads Failed to download this history item. Please try again.

Step 6: Rename or delete an item

Press Rename, or double-click the title, type a new name of up to 120 characters and press Enter. Press Escape to cancel. The page shows Renamed. If the name was not saved, the row reads The new name was not saved. Try again.

Press Delete, then Delete again where the row asks Delete for good?, or Keep to cancel. The page shows Deleted. Deletion is final. An item that is still Processing cannot be deleted until it finishes. Only the person who made an item, or a workspace admin, can delete it.

Step 7: Find your cloned voices

Open the Voice Cloning tab. Each cloned voice has an orb that plays its preview, its name, whether it is a professional clone, and the date it was made. Open takes you to the voice cloning page for that voice, Use in Text to Speech opens Text to Speech with it selected, and Delete removes it the same way as above. When there are none, the tab reads No cloned voices yet and Clone your voice and it lands here.

Worked example

On Monday you generated three versions of a product line, and on Tuesday you transcribed a 12-minute call.

  1. Open Assets. All shows four rows, newest first, under Today and Yesterday: the call, then the three generations, each with the line you wrote and a voice name.
  2. Type launch in the search box. Only the items whose text contains that word remain.
  3. Press Play on the second one. The button spins for a moment, then the player opens and the audio plays. Press Download: your downloads folder now holds a .wav file named after the first words of the line.
  4. Clear the search and press Rename on the call. Type Call with Acme and press Enter. Open the Speech to Text tab, press Download on it and pick SRT: Call with Acme.srt holds the whole transcript as subtitles, each cue starting with its speaker.

Nothing in these steps was charged.

What happens next

  • A generation or transcript made after the page was opened appears at the top the next time you open the page, or when you reload it.
  • An item still running shows Processing. Reload the page to see it change to Ready or Failed.
  • Items stay listed until you delete them, Failed ones included. Deletion is final.
  • Generations and transcripts made with an API key are not listed here. They are under the key's request log; see Request logs.
  • Cleaned recordings are listed here too, under Sound Cleaning. The newest 8 also show under Recent cleanups on the Sound Cleaning page; see Clean a recording.

Limits

Limit Value Applies to
What is listed Every generation, transcript and cleaned recording made from the portal in the workspace The Assets page
How many The newest 100 of each kind The tabs and the search
Name length 120 characters Rename
Removing an item Delete; final, the item leaves the workspace at once The Assets page
Time limit None; items stay until you delete them The Assets page

Troubleshooting

The page reads No history yet, but the workspace has generations

The list could not be loaded, or you are in another workspace. Reload the page. If it still reads No history yet, check which workspace you are in.

A generation or transcript I just made is missing

The list is loaded when the page opens. Reload the page; the new item is then at the top.

Failed to download this history item. Please try again.

The audio of a generation could not be fetched. Wait a moment and press Download again. If the row reads Processing or Failed, there is no file to download.

Play does nothing, or stops at once

The page has been open for a long time and the item's audio can no longer be reached. Reload the page and press Play again.

Assets opened on All when I came from the Speech to Text page

The page always opens on All. Click the Speech to Text tab.

Download is greyed out

The row reads Failed or Processing. A failed item has no file; a processing item has none until it finishes, so reload the page later.

FAQ

Who can see the list?

Every member of the workspace. The list is shared, not kept per person.

How long do items stay?

Until you delete them. Nothing is removed after a set time.

Can I delete an item?

Yes. Press Delete on the item, then Delete again where it asks Delete for good?. Deletion is final. Only the person who made the item, or a workspace admin, can delete it.

Does playing or downloading cost credits?

No. Playing, downloading, renaming and deleting are free.

Why is a generation made with my API key not here?

The page lists what was made from the portal. Generations and transcriptions made over the API are under Request logs.

Where are my cleaned recordings?

On the Sound Cleaning tab here, and under Recent cleanups on the Sound Cleaning page, which keeps the newest 8. See Clean a recording.