If you've ever tried to send a voice note on LinkedIn or WhatsApp through code, you know it's a nightmare. Every platform has different audio format requirements, different auth flows, and almost none of them have a documented public API for voice messages.
I spent months figuring this out for each platform. Here's what I learned.
The problem
Each platform handles voice notes differently:
LinkedIn expects M4A audio uploaded through an internal API, delivered with specific audio metadata and a waveform. There's no public API for this.
Telegram is the easiest. Their Bot API has a sendVoice method that accepts OGG Opus files.
WhatsApp requires a session-based connection and OGG Opus format.
Messenger, Instagram, Slack, Discord, Viber, WeChat all have their own quirks.
If you're building an outreach tool, a CRM, or any product that sends messages across platforms, supporting voice notes means dealing with all of these individually.
The audio format problem
Your user records an MP3. LinkedIn needs M4A. Telegram needs OGG Opus. WhatsApp needs OGG Opus but with different encoding parameters.
You end up building an audio conversion pipeline before you've even sent a single message.
What I built
I built Svara (https://svarapi.io) to abstract all of this into one API call:
bashcurl -X POST https://svarapi.io/api/v1/send \
-H "Authorization: Bearer sk_live_..." \
-H "Content-Type: application/json" \
-d '{
"platform": "telegram",
"recipient": "400533598",
"audio_url": "https://example.com/voice.mp3",
"session": {
"bot_token": "7000000000:AAH-abc123"
}
}'
You send any audio format. Svara converts it and delivers it as a native voice note on whichever platform you specify. The recipient sees a normal playable voice message, not a file attachment or a link.
Supported platforms
LinkedIn, Telegram, WhatsApp, Messenger, Instagram, Slack, Discord, Viber, and WeChat. Same endpoint, same request format.
If you want to try it
50 free voice notes at svarapi.io, no credit card required. API key in 2 minutes.
Would love to hear from anyone else who's dealt with the pain of multi-platform messaging APIs. What platforms are the worst to work with?
Top comments (0)