Voice dictation is always the same: you speak, the software turns your words into text, and then you spend time fixing typos and phrasing. Well, startup Relay introduced a different approach at the IFA 2026 with Relay Q, a dedicated push-to-talk hardware controller that understands real-time speech through context and not plain text transcription.
Launching first on macOS, the system analyzes raw spoken audio directly through a multimodal AI model. By taking active screen details into account—like open apps, selected text, and personalized memory—Relay can make sense of incomplete thoughts, mid-sentence self-corrections, and natural hesitation. Then it can output polished text or take direct action without making you touch the keyboard.
Modular hardware built around physical control
Relay Q doesn’t rely on an always-on ambient microphone. Instead of that, it centers user control around tactile physical inputs. Holding the main button records audio, releasing it generates the output, and double-clicking sends the result directly to your active app.
The hardware itself adopts a modular design. In handheld mode, the microphone tip attaches magnetically to its base to form a compact desk controller. When you need to move around or work hands-free, the tip detaches to function as a magnetic clip that attaches easily to a shirt collar, keeping the microphone close to your mouth for clear, quiet voice capture in busy environments.
Automating tasks and handling multilingual workflows
Beyond generating text, the platform expands its utility through Relay Skills. Users can verbally prompt the system to locate contacts, route messages through supported messaging apps, or fetch external web data—such as pulling a restaurant address directly into an active chat—without leaving their main window. Power users can also build custom skills to connect the interface with internal databases and task-management platforms.
The software is also built for multilingual users, allowing people to switch naturally between languages while speaking. Relay processes the mixed audio input and formats the final output into the language and tone appropriate for the intended recipient.
Pricing, privacy safeguards, and platform availability
Relay stores context data and memory locally on the user’s device, capturing audio only when the physical recording button is engaged. Data in transit is encrypted and never retained on external servers for model training.
Relay for macOS is available now with a 14-day free trial. Its Unlimited subscription tier costs $15 per month for Apple Silicon Macs. The company is also offering a $150 Early Supporter Bundle—available in Chrome, Pastel, or Brass finishes—which includes a pre-order for the Relay Q hardware, a magnetic charging dock, a MagSafe Remote, and a full year of software access ahead of its scheduled January 2027 hardware shipment.
Lastly, support for Windows, iOS, and Android is currently in development for release by early 2027.
The post Meet Relay Q: The Modular Push-to-Talk Gadget Turning Spoken Audio into Actual Action appeared first on Android Headlines.
​Â
