Overview
If a meeting didn't arrive in Alchemer Voice automatically, you can upload its recording. An upload creates a draft conversation. You add a transcript (import one or have Alchemer Voice transcribe the audio), confirm the details, and select Submit for analysis. Alchemer Voice generates insights only after you submit.
You need an active plan. A transcript can't be uploaded on its own: the recording comes first, and the transcript is added to the draft.
Supported files
| Item | Accepted |
|---|---|
| Recording | MP4 video; M4A or MP3 audio. Other types are ignored. Some browsers report MP3 files with a different type, in which case they may be skipped; convert to MP4 or M4A. |
| Transcript (optional) | WebVTT (.vtt), which most meeting tools export (including Microsoft Teams, Google Meet, and Dovetail). Also .txt, either WebVTT-style or one line per utterance in the form [00:01:23] Speaker name: what they said. Lines without a speaker name are skipped. |
No file size limit is shown on the upload page. Keep the page open until each progress bar completes.
Upload the recording
On Conversations, select Upload and drag one or more files onto the page. The draft's title starts as the file name. Drafts remain under Conversations > Draft conversations.
Add a transcript
Import an existing transcript
Drag the file onto "Drag and drop a transcript here". If it can't be read, you see "This transcript could not be imported"; check the format and retry.
Transcribe automatically
- Select Automatically Transcribe and choose the number of speakers (2, 3, 4, 5+, or Unknown). A correct count improves attribution; with five or more speakers, attribution can be unreliable.
- While "Transcription in progress" shows, you can leave and return.
- Speakers are named "Speaker 1", "Speaker 2", and so on until you rename them.
If you see "Transcription failed", delete the conversation and upload the file again. To replace a wrong transcript, choose Delete transcript from the draft's three-dot menu; the recording stays.
Confirm details and submit
Once a transcript exists, the Conversation details panel opens.
- Check the date. It defaults to the upload time.
- Choose a Purpose (for example Sales, Support, Product feedback, or Customer success). A purpose is required.
- Enter each participant's Email. It determines whether someone counts as internal or external (based on the internal email domains under Admin > Company), links them to CRM records, and affects which participants generate insights.
- Select Submit for analysis.
The conversation leaves Drafts and is processed. When finished, it appears under All conversations, and under My conversations if you're a participant. If your organization has no conversation analysis hours left, it shows as Not processed under Pending conversations. You can edit the date, purpose, and participants after submitting; see Conversations.
