ThunderPhone is a self-serve platform for building AI phone agents. From 2¢/min with models included. Our Storm tier, with its extra-intelligence option, scores 99.4% on Big Bench Audio — eval set public on Hugging Face. Build agents, test them with AI-caller simulations, monitor and steer live calls, and get auto-detected issues with proposed fixes. 47 languages.
Hi Product Hunt! Alex here, founder of ThunderPhone.
We spent the last year automating phone calls for enterprise customers and kept seeing the same failure: AI agents that mishear callers — names, addresses, spellings — until the caller gives up. So we built a different stack: a web of models where multiple transcription models and LLMs work each conversation turn together, instead of the usual three-step pipeline.
Last week we opened it to everyone: from 2¢/min with models included, an agent on a real phone number in minutes, AI-caller simulations to test it before it talks to a human, live call monitoring, and automatic issue detection on production calls. Our Storm tier (with the extra-intelligence option) scores 99.4% on Big Bench Audio — the eval set and transcripts are public on Hugging Face so you can check our work.
Try the live demo on our homepage without signing up. I'll be here all day — questions, feedback, and "it broke when I…" reports all welcome.
Report
@kolchinski How is the latency holding up with multiple models running consensus on every turn?
Mezli
@kolchinski How is the latency holding up with multiple models running consensus on every turn?