Sending voice messages

Voice notes are the highest-signal message format on LinkedIn: almost nobody sends them, so they get listened to.

Use send_voice_message with:

• audio_url (required): a public HTTPS URL to a .mp3, .m4a, .ogg, or .wav file, max 5MB
• chat_id for an existing conversation, or provider_id to start a new one
• text_caption (optional): text shown alongside the audio so the recipient has context at a glance
• campaign_id (optional): campaign attribution

The audio arrives as a native LinkedIn voice note, a playable waveform in their inbox, not a file attachment. Each voice note counts against your daily messaging limit (150 by default), and block-listed recipients are refused before anything sends.

Getting a hosted URL: if your audio file is local, upload it with upload_media, which returns a public URL ready to use. Any public HTTPS URL works too. For the audio itself, record it yourself or generate it with any text-to-speech provider, Crispy is deliberately agnostic about where the audio comes from.

Voice notes also work inside campaigns: a follow-up step with type voice_note and a media URL sends the note automatically after a connection accepts.

Common mistakes: - A URL that isn't publicly reachable. The send fails if the audio can't be fetched. - Wrong format or oversized file. Stick to the four supported formats and 5MB. - Sending the identical clip to hundreds of people. The format works because it feels personal, keep it personal.

Pro tip: pair every voice note with a one-line text_caption summarizing the ask. People often read before they listen, and the caption earns the listen.