Live Transcription

Live Transcription fills a transcript while people are still talking, instead of at the end of a recording.

How to use

Open Live Transcription in the sidebar and click New Recording
Set Speech Language
Click Start Recording
Pause suspends capture, Resume picks it up again, Stop & Save ends the session
In the Save Transcription dialog, fill in Title and save. The app hints "Tip: Configure post-processing in Settings to auto-generate titles" — with post-processing configured, a title is suggested for you

No API key required

Live Transcription is not ElevenLabs-only, and nothing here needs a BYOK key. Speakly picks the path from the engine set in Settings → Transcription. With ElevenLabs it opens a realtime stream: partial text updates while you speak and is committed on pauses. With every other engine — Groq, OpenAI, Deepgram, Mistral, Google, a custom endpoint, or the on-device Local Model — it runs a chunked path: voice activity detection cuts each phrase at your pauses and transcribes it immediately after. On the local model that means live transcription with no key and no network.

Two tracks: your mic and the system audio

Speakly can transcribe what your computer is playing — the other participants in a meeting, a video — as a second, fully independent track. That is how you get both sides of a call: your microphone is one track, everything coming out of your speakers is the other. The two tracks run their own voice activity detection and their own transcription requests in parallel, so a loud remote speaker cannot drown out your microphone.

Turning it on

Open Live Transcription → New Recording, but do not press Start Recording yet
Switch on Also capture system audio — "What your computer plays (participants, videos) is transcribed as its own track"
Press Start Recording. The switch is disabled while a session is running or paused, so it has to be set beforehand
The choice is remembered between sessions (stored as the liveCaptureSystemAudio setting)

The row only appears when system-audio capture is available on your platform and the engine is not the ElevenLabs realtime stream.

Platform support and permissions

macOS: a bundled ScreenCaptureKit helper captures the system output, so macOS asks for Screen Recording permission the first time. If it is denied, the session continues with the microphone only and Speakly shows "System audio unavailable (…) — recording microphone only. Check Screen Recording permission in System Settings → Privacy & Security."
Windows: WASAPI loopback through the app's display-media handler — no helper binary and no separate permission prompt
Linux: not supported; the toggle does not appear

What the transcript looks like

In a dual-track session, one line per phrase, prefixed with the wall-clock time it was spoken, e.g. [14:32:07]. A microphone-only session is one flowing paragraph with no prefixes.
On screen the timestamp is colour-coded by source: purple for your microphone, blue for the system audio — as the page's own footer puts it, "purple timestamps are your mic, blue ones are the system audio"
A line appears as … the moment speech starts and fills in when its text arrives, so lines never reorder when one track answers slower than the other
The saved dual-track transcript names the speaker in the text itself: [14:32:07] You: … for your microphone, [14:32:07] Them: … for the system audio. The colour exists only on screen, so writing the speaker into the line is what makes attribution survive saving and copying.

Muting one track

During a dual-track session the header shows a level meter per track. Click the microphone icon to mute your mic, or the speaker icon to mute the system audio: capture stays up, but nothing from that track is transcribed — or billed — until you unmute. Useful when you only want the other participants, or only yourself.

Meeting notes

Open a saved dual-track session from the list and the footer offers Meeting notes, whose tooltip reads "Summary, decisions and action items, attributed to each side". Because the transcript says who spoke, the output can too.

Summary — three to six sentences on what the meeting was about and where it landed
Decisions — one bullet each, saying who decided; "Nothing was decided." when nothing was
Action items — one bullet each as You — … or Them — …, with the deadline if one was said; "No action items." when there are none
Open questions — what was raised and left unresolved; "None." when there is nothing

The button only appears where the attribution does

Meeting notes is offered only for transcripts whose lines carry You: or Them:, and only when a post-processing provider is configured (the AI actions also need Pro — otherwise they are disabled with "Upgrade to Pro to use AI features"). Sessions saved before the speaker was written into the text have timestamps but no names, cannot be attributed after the fact, and do not get the button.

What you get

Live text: partial updates on the ElevenLabs stream, phrase by phrase on the chunked path
Level meters driven by the real captured signal, one per track
Per-track mute in dual-track sessions
Copy to clipboard and save to history
AI actions on a saved transcription: Flashcards, FAQ, Ask, Translate
Meeting notes on a saved dual-track session: summary, decisions, action items per side, open questions