Skip to main content

Team Radio

Latest team radio clip and rolling history (Replay Mode or F1TV Auth live timing). Use sensor.f1_team_radio in dashboards, templates, and automations.

See live data availability for session and data-source requirements. The details below describe any additional conditions for this entity.

Experimental transcription

Team radio transcription is experimental. It is tested with a Google Gemini AI Task entity, may change without notice, and sends audio to the AI provider only when you manually run the action. You are responsible for the provider API key, its terms, and any charges.

Find your entity

These are standard entity IDs. If Home Assistant assigned a different ID, or you renamed an entity, use your existing ID in the examples.

State and attributes​

Replay Mode or F1TV Auth live timing

This entity stays registered in Home Assistant. It updates in Replay Mode and can update during live sessions when F1TV Auth is paired with a valid token and live team radio data is available.

Latest curated team radio clip with a rolling history. This is not the full raw driver radio feed; it follows the Team Radio clips Formula 1 publishes for the session.

State

  • ISO-8601 timestamp of the most recent radio clip, or unknown when none are available.

Example

2026-07-05T14:01:01+00:00

Attributes

AttributeTypeDescription
utcstringISO-8601 timestamp of the radio clip
received_atstringISO-8601 timestamp when Home Assistant received the message
racing_numberstringCar number for the driver
pathstringRelative path to the audio file
clip_urlstringFull URL to the audio clip when available
sequencenumberMessage counter for deduplication
transcription_statusstringprocessing, completed, or failed after the transcription action is used
transcription_textstringComplete English transcript when available
transcription_segmentslistSpoken turns labeled anonymously as speaker A, B, or unknown
transcription_ai_task_entity_idstringAI Task entity that produced the transcript
transcription_generated_atstringISO-8601 timestamp when transcription completed or failed
transcription_errorstringShort error message when transcription failed
raw_messageobjectRaw team radio capture from the timing feed
historylistRolling list of recent radio clips, up to 20 items

Each entry in history contains:

FieldTypeDescription
utcstringISO-8601 timestamp of the radio clip
racing_numberstringCar number for the driver
pathstringRelative path to the audio file
clip_urlstringFull URL to the audio clip when available

Experimental transcription​

The experimental f1_sensor.transcribe_team_radio action sends the latest published clip to an AI Task entity. It keeps the original English language and separates distinct voices into anonymous speakers A and B. It does not identify the driver or engineer.

Before you start​

  1. Create a Gemini API key in Google AI Studio.
  2. In Home Assistant, open Settings > Devices & services > Add integration and add Google Gemini.
  3. Keep the recommended setup or create an AI Task subentry. Home Assistant then creates an entity similar to ai_task.google_ai_task.
  4. Wait for a live or replay team radio clip. Your team radio entity must have a clip_url attribute before it can be transcribed.

Only Google Gemini AI Task is supported for this experimental implementation. Other AI Task providers can be selected in the interface, but they have not been validated for MP3 transcription.

Transcribe the latest clip​

Open Developer Tools > Actions, select Experimental: Transcribe team radio, and enter your own entity IDs:

action: f1_sensor.transcribe_team_radio
data:
  entity_id: sensor.f1_team_radio
  ai_task_entity_id: ai_task.google_ai_task

The status changes to processing while the request runs and then to completed or failed. If it fails, read transcription_error on the team radio entity.

View the result​

The complete transcript is stored in transcription_text. Individual speaker turns are available in transcription_segments, where speakers are labeled A, B, or unknown.

The structured speaker turns can be displayed in a Markdown card:

type: markdown
content: |-
  {% for segment in state_attr('sensor.f1_team_radio', 'transcription_segments') or [] %}
  **{{ segment.speaker }}:** {{ segment.text }}

  {% endfor %}

Cost, privacy, and limitations​

The action is manual. F1 Sensor never sends team radio audio to an AI provider automatically. When you run it, the MP3 file is sent to the selected provider and may incur charges. Temporary local audio files are removed after the request.

A completed result is cached for the current clip while F1 Sensor remains loaded. Running the action again with force: false does not create another AI request for that clip. force: true deliberately sends a new request. The cache resets when the integration reloads or Home Assistant restarts.

Troubleshooting​

  • No team radio clip is currently available: wait for a new live or replay radio clip and confirm that clip_url is present.
  • Select an AI Task entity: choose the Google Gemini AI Task entity, not its conversation or speech-to-text entity.
  • Transcription failed: check transcription_error, then confirm that the Gemini API key is valid and that the configured account can use the selected model.

Next steps​