`summarize transcriber`

summarize transcriber setup [--model parakeet|canary] [--theme <name>]

Configures local ONNX transcription. The command prints the env vars you need to export — it doesn't write to your shell config. Once those vars are exported, --transcriber auto (or SUMMARIZE_TRANSCRIBER=auto) prefers the local path.

#How auto picks a transcriber

When --transcriber auto is set (default), summarize tries providers in this order:

Groq — fastest cloud option. Needs GROQ_API_KEY.
Local ONNX (Parakeet or Canary) or whisper.cpp — fully offline. Needs the env vars from this command.
AssemblyAI / Gemini / OpenAI / FAL — cloud fallbacks, in that order, gated by their respective API keys.

Skipping all of those falls back to the OpenAI-compatible Whisper endpoint in OPENAI_WHISPER_BASE_URL.

#Subcommand

#`summarize transcriber setup`

--model <name> : parakeet (default) or canary. Picks which ONNX model the printed snippet targets.

--theme <name> : CLI theme override.

The output is a labelled, copy-pasteable block of export lines for your shell. Re-run with the other model to switch.

#Environment

Set by the printed snippet:

SUMMARIZE_ONNX_PARAKEET_CMD : Command to run Parakeet ONNX transcription. Use {input} as the audio-file placeholder.

SUMMARIZE_ONNX_CANARY_CMD : Command to run Canary ONNX transcription. Use {input} as the audio-file placeholder.

SUMMARIZE_TRANSCRIBER : Optional. Lock in auto, whisper, parakeet, or canary without passing --transcriber every time.

#Examples

# Default: Parakeet.
summarize transcriber setup

# Set up Canary instead.
summarize transcriber setup --model canary

# After exporting the printed vars:
summarize "https://podcasts.apple.com/.../episode-..." --transcriber parakeet

#See also

ONNX transcription — how to actually install the model + runtime.
Transcript provider flow — full selection waterfall.

summarize transcriber