Who does it compete with for the same buyer, and how do they differ?
For realtime agent companies, Hopper's direct alternatives are voice-model vendors such as ElevenLabs, Cartesia, and Deepgram, which sell their own voice-agent products that directly compete with their customers, while Hopper works exclusively with realtime agent companies.
Hopper prices Qwen 3 TTS at 150 ms TTFA as 10x less expensive than ElevenLabs, and prices Nemotron 3.5 ASR as 2x faster and 8x less expensive than Deepgram Nova-3.
2Pricing Strategy
How does its pricing position it against rivals?
Hopper offers voice-model serving below a named frontier alternative, namely the $2 rate for GPT-4.1.
The Starter plan for Hopper Code costs $20 per month for everyday coding, the Pro plan costs $50 per month for all-day coding, and the Max plan costs $200 per month for parallel agents.
3Product Overview
What does the product do for its buyer, and how is it delivered?
Hopper runs managed inference on open-weight speech models for realtime agent companies, updating those models with live customer traffic and serving exclusively that buyer segment.
Hopper hosts its cloud infrastructure across the United States, the European Union, and India, and partners with each customer's engineers to co-design the initial evaluation.
Hopper
Hopper runs open-weight speech-model inference that updates on live customer traffic.