Okay, so I've been deep in the weeds on a migration project for our sales team's meeting intelligence pipeline. We were using Granola for a while, but the pricing jump on the new tiers has us looking around.
I love Granola's focus on developer-friendly APIs and webhook workflows—it fit nicely into our existing DevOps tooling. But now we need to evaluate alternatives. The usual suspects everyone name-drops are Otter.ai and Fireflies.ai, but honestly, they feel like a step backward for us? Too much focus on the end-user interface, not enough on the raw data pipeline and flexibility.
I'm hunting for something that can:
* Handle the transcription/recording ingestion reliably (obviously).
* Give me clean, webhook-driven events for "meeting processed" or "summary ready."
* Expose a decent API for pulling transcripts and structured data (speakers, topics) into our own data warehouse.
* Not break the bank for ~500 meetings/month.
What I'm **not** looking for is just another chatbot-in-your-meeting app. I want the *engine*, not the shiny car.
So, besides the two big names in the title... what's out there? Has anyone tried:
- **Assembly.ai** for this use case? Their API looks promising, but it's more generic speech-to-text.
- **Dovetail**? I know it's big for research, but can it pipe meeting data *out* easily?
- Any lesser-known SaaS that acts more like "Granola-as-a-service" where the output is a JSON blob in my S3 bucket?
My nightmare scenario is getting locked into a platform where the only export is a CSV from a UI. I need to automate everything. War stories and gotchas are especially welcome 🙏 What broke during your POC, and how did you fix it?
Assembly's API is solid, but you're paying for general speech-to-text accuracy, not the meeting-specific intel like speaker diarization baked into your pipeline. Their model is a raw material, not a finished part.
Have you looked at Rev's APIs? They're utilitarian but reliable, and the webhooks for job completion are exactly what you're after. The cost scales cleanly per minute, no platform bloat.
But the real question is why you're buying this as a service at all. For 500 meetings, you could pipe recordings to Whisper via your own script and trigger events off the output for half the cost. The "engine" you want is just a cloud function away.