> ## Documentation Index
> Fetch the complete documentation index at: https://docs.weelkit.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Transcription

> Record a conversation, upload an audio file or turn a public video link into a transcript with speakers, then rename speakers, copy, download or save it.

Open **Tools → Transcription**. It is listed on [All Tools](https://docs.weelkit.com/images/tools.md) while your organization has the tool turned on; an admin turns it on in [Organization settings → Tools](https://docs.weelkit.com/settings/organization.md#tools).

## Record a conversation

1. Choose **Start recording**, and allow the microphone.
2. Choose **Pause recording** to pause, and **Resume recording** to carry on. Paused time doesn't count towards the limit.
3. Choose **Stop recording**. The transcript appears under **Transcriptions**.

A recording can be up to four hours long. Tell everyone in the room that you are recording before you start: recording a conversation without the knowledge of the people in it may be unlawful where you are. The panel says this above the controls.

### Close the tab without losing the recording

Your recording is saved on this device as you record and uploads while the page is open. If you close the tab or lose the connection, your next visit lists the recording under **Unfinished recordings**. Choose **Upload** to finish it, or **Finish** when everything is already uploaded. **Discard** throws the recording away after asking.

If you try to leave mid-recording, Weelkit asks **Leave while recording?** Choose **Keep recording** to stay, or **Stop and leave** to stop the recording. The audio already saved stays on this device.

> [!NOTE]
>
> If you see **This recording is in memory only**, don't refresh the page: you would lose the recording. If you see **Keep this tab open**, keep the tab open until the upload finishes.

## Upload an audio file

1. Choose **Choose audio file**. One file at a time: MP3, M4A, WAV, OGG or WebM audio.
2. Choose **Transcribe audio**.

A video file isn't accepted here. Paste its link below instead: only the audio track is transcribed.

## Transcribe a video link

1. Paste the address under **Video link**. Links only — a video file isn't accepted.
2. Choose **Transcribe link**.

The link must be public. Private or sign-in-only videos can't be transcribed. Every link transcript is speaker-labelled.

If video links aren't available in your organization, Weelkit says so. Upload the audio instead.

## Read a transcript

**Transcriptions** lists every job with its source (**Audio file**, **Recording**, **Video link** or **Chat attachment**), its length, its language and its state:

| State | Means |
| - | - |
| **Queued** | Waiting to start. |
| **Transcribing** | In progress. Long recordings can take a while. |
| **Done** | The transcript is ready. Choose **Open** to read it. |
| **Failed** | The audio couldn't be transcribed; the row says why. |
| **Outcome unknown** | Weelkit couldn't confirm the result. The transcript page says why. Start it again if you still need the transcript. |

The transcript shows:

* **Speakers:** choose **Rename** beside a speaker, type the name, then choose **Save name**. **Cancel** leaves the name as it was.
* **Transcript:** each line shows when it was said, then the speaker and the text.
* **Copy:** copies the whole transcript, with speaker names and timestamps.
* **Download text** and **Download Markdown:** download the transcript as a `.txt` or `.md` file.
* **Save to library:** saves the transcript as a Reference in [Knowledge](https://docs.weelkit.com/library-and-knowledge/knowledge.md), where your organization can read it and the AI can use it. After it saves, the page says **Saved to the library**.
* **Delete:** removes the audio and the transcript after you confirm. This can't be undone.

Choose **Back to transcriptions** to return to the list.

## Use a transcript in chat

* **An audio file:** choose **+**, then **Add files or photos**, and attach an audio file. Ask the AI to transcribe it.
* **A video link:** paste the link in your message and ask the AI to transcribe it. Every link transcript is speaker-labelled.

A long recording or video can take a while: the AI tells you it is running, and the transcript arrives in the chat when it's ready. Chat work appears in **Transcriptions** too, with **Chat attachment** or **Video link** as its source.

> [!NOTE]
>
> AI can make mistakes. Check names, figures and quotes before you rely on a transcript.

## If something goes wrong

* **This tool is disabled for your organization.** The tool is off. An admin can turn it on in [Organization settings → Tools](https://docs.weelkit.com/settings/organization.md#tools).
* **The upload didn't finish.** The recording stays on this device. Choose **Upload** on the unfinished recording to try again.
* **The transcriptions could not be loaded.** Choose **Try again** to read the list again.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.