Your Dataset lives on the Dataset tab under Simulations. Adding a call from production is free. Uploads are billed per minute of audio.
Add a call
Open the dialog, name the call, choose a source, and extract it. Only step 3 depends on your source.1
Open the dialog
Go to Simulations > Dataset and click Add call.
2
Name the call
Enter a Name: required, up to 255 characters, and unique for this agent. Add a Description if you like, up to 1,000 characters, saying what makes the call worth replaying. It shows on the call’s card.
3
Choose a source
Choose Production for a call Tuner already recorded from your agent, or Upload for an audio file you already have.
- Production
- Upload
Once your voice agent is connected to Tuner, the real calls it takes show up in Tuner with a recording and a transcript. With this option, you pick one of those calls and add it to your Dataset. It’s the quickest way to save a call that went wrong, because everything Tuner needs is already there.

- Cost: free.
- Click the call you want from the list. The newest calls come first, and you can search by call ID.
- Only calls that have both a recording and a transcript are listed. A recording link counts only if it’s a public
https://URL. Calls sent with a placeholder link, such aspending, aren’t listed. Simulation test calls aren’t listed either, but you can save one as a scenario and run it again instead. - Tuner already knows which side is the caller, so there’s nothing to pick.
- Prefer calls marked Multi-channel, where the caller and your agent are recorded separately. On a Mono source call, the agent’s voice can bleed into the caller’s turns.
4
Extract the call
Optional: expand Advanced extraction to change how the caller’s speech is cut (see Advanced extraction). Then click Extract call.
5
Wait for Ready
The call appears on the Dataset tab as Extracting, which updates on its own. It’s usually Ready within a minute, and then it can be replayed. If it shows Failed, see Troubleshooting.
After you add a call
Check the status
Review the caller turns
Click the headphones icon on a Ready call to open its turns. Each turn shows its position in the original recording, its duration, its reference transcript, and whether that transcript is Draft or Verified. Click play to hear just that turn. For production calls, View source call opens the original call.
- Clipped words at the start or end of a turn: raise Padding and re-extract.
- Agent speech inside a caller turn: common with mono sources. Pick a multi-channel call if you can.
- Turns that are split or merged in the wrong place: adjust Merge gap or Minimum turn and re-extract.
Edit and verify transcripts
Tuner writes a draft reference transcript for every caller turn. When a call is replayed, Tuner compares these transcripts with what your agent heard to score transcription accuracy. Check that each transcript matches what the caller said. A wrong reference transcript makes the score wrong.1
Open the turns
Click the headphones icon on a Ready call. The header shows how many turns are verified, like 7 of 10 verified. Unverified turns are listed first.

2
Listen and read
Click play on a turn and read its transcript. To see only some turns, use the filter: All, Draft or Verified.
3
Verify or correct it
If the transcript is right, click Verify. If it needs changes, click Edit, change the text, and click Save and verify. Click Cancel to discard your change.

- Saving an edit always marks the turn Verified. To undo that, click Unverify on the turn.
- A transcript can be empty, for a turn where the caller said nothing.
- Editing or verifying never recalculates earlier replays. Open the call and click Recalculate. See Outdated scores.
- Only Ready calls can be edited.
Re-extract a call
If the turns come out wrong, cut the call again with different settings. Click the refresh icon on a production call. Tuner creates a new Dataset call from the same recording and suggests a versioned name, like Billing dispute, Feb 4 v2. The original stays as it is, so earlier replays remain comparable. Your edits stay on the original. Every turn of the new call starts as Draft.- The extraction settings start from the defaults, so set them again before you extract.
- Uploaded calls can’t be re-extracted. Upload the file again instead.
Advanced extraction
These settings are in the Add call dialog, and you set them again when you re-extract. The defaults suit most calls.
Merge gap isn’t used for uploads, or for multi-channel LiveKit, Pipecat, Dograh, and Custom API calls, which merge automatically unless the agent speaks in between. Padding is always 0 for mono sources.
Delete a call
Click the trash icon and confirm. The call can no longer be replayed, but runs that already replayed it keep their results. Its name becomes available again.Add calls with the API
You can also build your Dataset from code, for example to add every escalated call as it happens. The endpoints are in the Replay testing group of the API Reference, which has the request details and examples.- From production
- From an upload
- Find the call with List Source Call Candidates.
- Create the Dataset call with Create Voice Clip Set.
- Poll Get Voice Clip Set until the call is ready or failed.
Limits
Troubleshooting
A call shows Failed
A call shows Failed
Nothing was extracted, and the reason shows on the call’s card. Common reasons:
An upload is rejected
An upload is rejected
The file is checked during analysis, before you extract.
Next steps
Replays
Replay your Dataset calls against your agent and read the results.
Dataset best practices
How to choose, name, and maintain calls so every replay tells you something.
