Voice to Chat / EMail would be the killer feature. When I read an email, I just want to give a 3-5 sec voice to tell how to answer. It's much faster than writing the prompt myself. Pls let me know.
in progress
J
John Hertenberger
I'd love for Merlin's voice input to function like ChatGPT's voice-to-text—specifically using OpenAI's Whisper model or similar technology, not like native device dictation (Apple/Google).
The ideal user experience would be:
Press to record:
- Speak naturally for 3-5 seconds
- -Tap to stop
- Allow a brief processing moment (2-3 seconds) for superior accuracy
- Get highly accurate, context-aware transcription
This approach, powered by Whisper (which is open-source and can be run on Merlin's own servers), delivers far better accuracy and understanding than real-time dictation. The small processing delay is worth it for the dramatically improved transcription quality. Whisper handles accents, background noise, and complex language remarkably well, making it perfect for professional environments.
I'd much prefer this high-accuracy approach, even with the brief processing time, over instant-but-less-accurate live transcription. If Merlin could implement this type of server-processed, AI-driven voice transcription (whether using Whisper or similar technology), it would be a game-changing feature for daily productivity! It's the main thing keeping me from switching over completely and renewing. Thank you!
Lakshya Kumar
Merged in a post:
voice input feature
N
Nagur Bhasha
Merlin AI currently lacks a voice input feature, unlike some other, Integrating voice search capability would enhance its usability and accessibility significantly, enabling users to interact more naturally and efficiently.
Lakshya Kumar
in progress
Currently work in progress, adding it to core chat experience first.
h
herrkaschel
Lakshya Kumar Thanks! Look into TalkTastic. That's what their app does. It screenshots when you begin dictating to learn the context and e.g. important names, to get them right in the output. No matter how unpolished your dictation is, if you make pauses, lots of umms, or diverge, it crafts a near-perfect output text and pastes it where the cursor currently lies.
Having this in Merlin would be fantastic!
E
Eco
This feature would be great, talking instead of typing. Please start recording with 1x = microphone icon and click once for = off or stop recording (also rework in the app version). Please not 4-5 seconds, but as long as prompts as you like. No car off as is currently the case in the app version.
endu
under review
endu
Thanks for sharing this idea...That’s really interesting, and I can see how that would save a lot of time. We’ll definitely look into it—really appreciate your input
endu
Merged in a post:
Voice input in the browser
M
Markus
For me, it would be very useful if voice input were possible when using Merlin in the browser. The same goes for the Chrome plugin.
Merlin
Hi Nagur ji. Currently, we have voice input on the Merlin Mobile App (on Android and iOS). We're in process of figuring out voice input on Web as well, to bring you the perfect voice experience. Rest assured, we're working actively to bring a legendary voice UX to Merlin. Thanks for your ideas!