Transcript and Playback

Once a recording is uploaded to the cloud, you can generate a “Transcript” for it: an AI speech model converts the entire recording to text, automatically splits it by speaker, and adds timestamps — and text stays in sync with audio while you listen. This article covers generating, browsing, and playing back a Transcript.

Generating a Transcript

1. In “My Recordings”, tap the target recording to open its details. (The recording must be uploaded to the cloud; if the top of the home page still shows “Uploading”, wait for it to finish — upload rules are in “Sync and Upload”.)
2. Go to the “Transcript” tab and tap “Generate Transcript”.
3. The recording file uploads to the server, where the AI speech model converts speech to text and separates speakers. The page shows “Generating transcript. You will be notified when done.” — you can leave anytime; processing continues in the background.
4. When it’s done, the “LanEcho” service account sends a card notification, “Transcript Generated”. Tap the card to go straight to that recording’s “Transcript” tab.

Good to know:

• Before a Transcript is generated, the top of the page shows “Remaining available transcription time this month: xx minutes”. The remaining quota depends on your organization’s licensing.
• If the task is queued, you’ll see “Waiting to generate transcript. You will be notified when done.” — no need to wait on the page.
• Recordings made with “Real-Time Transcription” get a Transcript automatically after upload — no manual tap needed.
• If quota is insufficient, you’ll see “Generation failed: no available quota. Please contact your organization administrator.” If generation fails, a “Transcript Generation Failed” notification arrives, and you can try generating again.
• If transcription times out or the recording file is corrupted, transcription can’t complete — the page explains why and offers a re-transcribe entry.
• For recordings over 2 hours, you’ll be told transcription takes longer and can run in the background, with the same notification when done.
• For recordings in “Shared With Me” without a Transcript, manual generation isn’t supported — tapping shows “No permission”. The recording owner needs to generate it before you can view.

Browsing a Transcript

Once generated, the “Transcript” tab presents everything segmented by speaker:

• Each speaker shows an avatar and name (“Speaker 1”, “Speaker 2”… by default). There are 10 built-in avatars, and the same speaker always keeps the same avatar for easy recognition.
• Every paragraph starts with a timestamp (hh:mm:ss, relative to the start of the recording), so you can see at a glance when something was said.
• Photos taken during the recording appear at their matching timestamps, corroborating the text.
• Marks dropped during recording (short-press the record button on the device, or the “Mark” button in the Remote — see “Recording”) show as orange markers.
• Long-press any text paragraph to copy it — handy for quoting.

A disclaimer sits at the end of every Transcript: “The above content is generated by an AI speech model and is for reference only.” Technical terms, dialects, and similar content may be misrecognized — double-check anything important by listening back.

Playing Back a Recording

The Transcript works in lockstep with audio playback. Combine the positioning methods below for faster review:

Playback Controls

• When the recording file is on your phone, tap the play button to start. The button area shows “elapsed time / total duration” with a pause button.
• When the file isn’t local, tapping downloads it first (“Downloading recording file” with progress), then playback begins. If the download fails, you’ll see “Failed to download. Please try again.”
• Sending Lansenger to the background doesn’t interrupt playback. Playback doesn’t loop when it ends — it stays in the finished state.
• Playback stops if you pause manually, playback finishes, you leave the recording details page, you exit LanEcho, or another app takes over the audio channel. Tap play again when you’re back.

Text–Audio Linked Positioning

• Paragraph highlighting: during playback, the current paragraph — down to the exact word — is highlighted in light blue, clearly distinct from the rest. When the highlight scrolls off screen, the page auto-scrolls to follow it, keeping the current position visible. If you scroll away to browse, the page returns to the current playback position after about 5 seconds of inactivity.
• Tap to play: tap any paragraph and playback starts from the nearest matching position, with the progress bar jumping in sync — perfect for “read first, then verify by ear”.
• Marker jumps: tap an orange marker in the Transcript to jump straight to the marked moment (no tone plays when you jump this way). When normal playback reaches a marked moment, a tone reminds you this is a highlight.

If playback says the recording file no longer exists (the local copy was deleted), re-download it as the dialog suggests — cloud data is unaffected.

Viewing Recording Info

The “Recording Info” tab in recording details shows the basics for verification and traceability:

• Creator: who made the recording; Created at: when the recording started.
• Synced at: when the recording synced to the phone; Duration: hh:mm:ss.
• File Size: size of the local recording file (MB); Source: the recording device name or the sharer’s name.

Loading on a Weak Network

Transcripts can be large. On a weak network (an underground garage, a meeting room with poor signal), opening the “Transcript” tab:

• The page shows a download progress bar (for example, “460KB / 2.53MB”) so you can watch loading progress instead of staring at a blank page. Switch to another tab and back mid-download and the progress is still there.
• If loading fails, it retries automatically 3 times. If it still fails, the page explains why (for example, “Your network is unstable. Try switching to Wi-Fi and retrying.”) and offers a manual retry.
• If automatic and manual retries both fail, loading is attempted again the next time you open the recording details — no need to regenerate the Transcript.

Next: Editing Speakers

Default names like “Speaker 1” and “Speaker 2” don’t read well. Rename the key participants to their real names: tap the speaker name before a paragraph to edit — you can pick people from your Lansenger contacts and replace names in batches. See “Editing Speaker Names”. Once that’s done, generate the Smart Summary for more accurate people info — see “Smart Summary and Chapters”.