ElevenLabs brings lifelike voice to the rabbit r1
ElevenLabs just announced a partnership with rabbit to bring its low-latency, emotionally rich voice models to the pocket-sized r1 companion. Users can simply press a button and speak to execute multi-level tasks like hailing rides and ordering food through rabbit OS. Why it matters: natural-sounding voice is the missing piece in making standalone AI hardware feel like a real assistant rather than a robotic gimmick.
That quirky orange AI gadget from CES is finally getting a proper voice.
ElevenLabs just announced a partnership with rabbit to bring lifelike speech to the r1 pocket companion. Instead of robotic, monotone responses, the little device will now talk back using ElevenLabs' low-latency, highly expressive voice models.
Giving rabbit r1 a real personality
The rabbit r1 is built around hands-free simplicity. Instead of tapping through dozens of smartphone apps, you push a single button and talk.
Behind the scenes, rabbit OS runs on a proprietary Large Action Model (LAM). It is engineered to carry out real-world digital tasks on your behalf, like booking an Uber, ordering dinner, generating Midjourney images, or scanning your fridge with its rotating camera to propose recipes.
To make that experience feel convincing, the audio feedback needs to feel human. rabbit founder and CEO Jesse Lyu explained that natural language only works if the interaction feels as effortless as talking to a real personal assistant.
That is where ElevenLabs comes in. Known for supporting 29 languages and crafting emotionally rich voices, the company has fine-tuned models specifically tailored for latency-sensitive conversational tasks.
As ElevenLabs CEO Mati Staniszewski put it, the goal is to turn the r1 into a truly dynamic co-pilot.
Why it matters
For anyone keeping an eye on the wave of dedicated AI gadgets, voice quality makes or breaks the entire product. If an assistant takes too long to reply or sounds like a GPS unit from 2008, people immediately stop using it. Giving the r1 a lifelike, low-latency voice removes huge friction from voice-first computing.
From an industry angle, this partnership shows how smart hardware startups are building their stacks. Rabbit is not trying to reinvent speech synthesis from scratch. They focused on industrial design and their action-oriented model, while outsourcing audio to the best player on the field.
We think it is a sharp tactical move, but real-world execution will be the ultimate test. A chain that requires on-device mic input, cloud-based LAM reasoning, cross-app execution, and streaming cloud TTS has plenty of places to lag. If the overall latency stays low, r1 could deliver one of the most natural companion experiences on the market.
Sources (1)
- elevenlabs.io(opens in a new tab)Originalelevenlabs.io