Skip to content

Add live transcription preview #571

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
basnijholt merged 11 commits into from
Jun 5, 2026
Merged

Add live transcription preview #571

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
8abad75
Add live transcription preview
basnijholt Jun 3, 2026
57c3326
Update auto-generated docs
github-actions[bot] Jun 3, 2026
e5ff5c1
Make live preview overlay opt-in
basnijholt Jun 4, 2026
204f561
Update auto-generated docs
github-actions[bot] Jun 4, 2026
a70fd1a
Document live transcription preview
basnijholt Jun 4, 2026
c97ee9a
Address live preview review feedback
basnijholt Jun 4, 2026
f41149c
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Jun 4, 2026
930e888
Fix live preview Swift test expectations
basnijholt Jun 4, 2026
5b2893b
Keep Rich status for inactive live preview console
basnijholt Jun 4, 2026
842a366
Cover live preview cancel before final race
basnijholt Jun 4, 2026
189f5b4
Signal live preview stop before cancellation
basnijholt Jun 5, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 15 additions & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -923,6 +923,21 @@ the `[defaults]` section of your configuration file.
│ provide context for │
│ LLM cleanup. │
╰────────────────────────────────────────────────────────────────────────────────────────╯
╭─ Live Preview ─────────────────────────────────────────────────────────────────────────╮
│ --live-preview-log PATH Write rolling live transcription │
│ preview events to JSONL while │
│ recording. │
│ --live-preview-interval FLOAT Seconds between live preview │
│ retranscriptions. │
│ [default: 2.0] │
│ --live-preview-window FLOAT Seconds of recent audio to include │
│ in each live preview │
│ retranscription. │
│ [default: 15.0] │
│ --live-preview-console,--live-previ… Print rolling live transcription │
│ preview updates to the terminal │
│ while recording. │
╰────────────────────────────────────────────────────────────────────────────────────────╯
╭─ Diarization ──────────────────────────────────────────────────────────────────────────╮
│ --diarize --no-diarize Enable speaker │
│ diarization │
Expand Down
51 changes: 44 additions & 7 deletions agent_cli/agents/transcribe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -365,13 +365,20 @@ async def _async_main( # noqa: PLR0912, PLR0915, C901
emit_output: bool = True,
raise_diarization_errors: bool = False,
audio_level_callback: Callable[[bytes], None] | None = None,
live_preview_log: Path | None = None,
live_preview_interval: float = 2.0,
live_preview_window: float = 15.0,
live_preview_console: bool = False,
) -> TranscriptResult:
"""Unified async entry point for both live and file-based transcription."""
start_time = time.monotonic()
transcript: str | None
saved_recording_path: Path | None = None
live_preview_console_active = (
live_preview_console and audio_file_path is None and provider_cfg.asr_provider == "wyoming"
)

with maybe_live(not general_cfg.quiet) as live:
with maybe_live(not general_cfg.quiet and not live_preview_console_active) as live:
if audio_file_path:
# File-based transcription
# Determine if we can use native format support (skip PCM conversion)
Expand Down Expand Up @@ -447,6 +454,17 @@ def _set_saved_recording_path(path: Path) -> None:
openai_asr_cfg,
gemini_asr_cfg,
)
live_preview_config = (
asr.LivePreviewConfig(
log_file=live_preview_log,
interval_seconds=live_preview_interval,
window_seconds=live_preview_window,
console=live_preview_console,
)
if (live_preview_log or live_preview_console)
and provider_cfg.asr_provider == "wyoming"
else None
)
transcript = await live_transcriber(
logger=LOGGER,
stop_event=stop_event,
Expand All @@ -456,6 +474,7 @@ def _set_saved_recording_path(path: Path) -> None:
extra_instructions=extra_instructions,
recording_path_callback=_set_saved_recording_path,
audio_level_callback=audio_level_callback,
live_preview_config=live_preview_config,
)

elapsed = time.monotonic() - start_time
Expand Down Expand Up @@ -650,6 +669,10 @@ def transcribe( # noqa: PLR0912, PLR0911, PLR0915, C901
print_args: bool = opts.PRINT_ARGS,
transcription_log: Path | None = opts.TRANSCRIPTION_LOG,
voice_level_log: Path | None = opts.VOICE_LEVEL_LOG,
live_preview_log: Path | None = opts.LIVE_PREVIEW_LOG,
live_preview_interval: float = opts.LIVE_PREVIEW_INTERVAL,
live_preview_window: float = opts.LIVE_PREVIEW_WINDOW,
live_preview_console: bool = opts.LIVE_PREVIEW_CONSOLE,
# --- Diarization Options ---
diarize: bool = opts.DIARIZE,
diarize_format: opts.DiarizeFormat = opts.DIARIZE_FORMAT,
Expand Down Expand Up @@ -697,18 +720,24 @@ def transcribe( # noqa: PLR0912, PLR0911, PLR0915, C901

setup_logging(log_level, log_file, quiet=effective_quiet)

enroll_speakers = _option_default(enroll_speakers)
identify_speakers = _option_default(identify_speakers)
remember_unknown_speakers = _option_default(remember_unknown_speakers)
speaker_profiles_file = _option_default(speaker_profiles_file)
speaker_match_threshold = _option_default(speaker_match_threshold)
live_preview_log = _option_default(live_preview_log)
live_preview_interval = _option_default(live_preview_interval)
live_preview_window = _option_default(live_preview_window)
live_preview_console = _option_default(live_preview_console)

# Expand user path for transcription log
if transcription_log:
transcription_log = transcription_log.expanduser()
voice_level_log = _option_default(voice_level_log)
if voice_level_log:
voice_level_log = voice_level_log.expanduser()

enroll_speakers = _option_default(enroll_speakers)
identify_speakers = _option_default(identify_speakers)
remember_unknown_speakers = _option_default(remember_unknown_speakers)
speaker_profiles_file = _option_default(speaker_profiles_file)
speaker_match_threshold = _option_default(speaker_match_threshold)
if live_preview_log:
live_preview_log = live_preview_log.expanduser()

# Validate diarization options
if not diarize and (enroll_speakers or remember_unknown_speakers):
Expand Down Expand Up @@ -850,6 +879,10 @@ def transcribe( # noqa: PLR0912, PLR0911, PLR0915, C901
diarization_cfg=diarization_cfg,
emit_output=not json_output,
raise_diarization_errors=diarize,
live_preview_log=live_preview_log,
live_preview_interval=live_preview_interval,
live_preview_window=live_preview_window,
live_preview_console=live_preview_console,
),
)
except ImportError as exc:
Expand Down Expand Up @@ -942,6 +975,10 @@ def transcribe( # noqa: PLR0912, PLR0911, PLR0915, C901
audio_level_callback=audio_level_writer.write_chunk
if audio_level_writer
else None,
live_preview_log=live_preview_log,
live_preview_interval=live_preview_interval,
live_preview_window=live_preview_window,
live_preview_console=live_preview_console,
),
)
except ImportError as exc:
Expand Down
25 changes: 25 additions & 0 deletions agent_cli/opts.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -436,6 +436,31 @@ def _conf_callback(ctx: typer.Context, param: typer.CallbackParam, value: str) -
hidden=True,
rich_help_panel="General Options",
)
LIVE_PREVIEW_LOG: Path | None = typer.Option(
None,
"--live-preview-log",
help="Write rolling live transcription preview events to JSONL while recording.",
rich_help_panel="Live Preview",
)
LIVE_PREVIEW_INTERVAL: float = typer.Option(
2.0,
"--live-preview-interval",
help="Seconds between live preview retranscriptions.",
rich_help_panel="Live Preview",
)
LIVE_PREVIEW_WINDOW: float = typer.Option(
15.0,
"--live-preview-window",
help="Seconds of recent audio to include in each live preview retranscription.",
rich_help_panel="Live Preview",
)
LIVE_PREVIEW_CONSOLE: bool = typer.Option(
False, # noqa: FBT003
"--live-preview-console",
"--live-preview-stdout",
help="Print rolling live transcription preview updates to the terminal while recording.",
rich_help_panel="Live Preview",
)

# --- Server Options ---
SERVER_HOST: str = typer.Option(
Expand Down
Loading
Loading