The Hedy Transcription Difference
When we set out to create Hedy, we knew traditional transcription methods fell short. Most apps record your audio and send it to distant servers, compromising your privacy. We took a radically different approach: Hedy processes speech locally on your device in real time. By default, audio is never saved – it exists only momentarily in your device's memory during processing and is immediately discarded afterward. Only the resulting text transcript is kept.
Local Processing: Your Conversations, Your Control
Hedy performs speech recognition entirely on your device. This isn't just a feature – it's a fundamental reimagining of how conversation technology should work. By default:
Audio is not stored. Hedy does not save recordings of your conversations. Audio data is held temporarily in RAM only for the brief moment needed to convert speech to text, then immediately discarded.
Audio is never sent to external servers. All speech processing happens locally on your device – no audio data is transmitted to Hedy's servers or any third party.
Only text is retained. What Hedy keeps is the resulting transcript – a text representation of the conversation, not the audio itself.
If you'd like to keep your audio for later review, you can enable audio saving in your session settings. When this option is turned on, Hedy stores the audio recording locally on your device alongside the transcript. Even with this option enabled, audio is never sent to external servers.
Privacy isn't an afterthought – it's our core principle. By processing audio locally and discarding it by default, we ensure your conversations remain completely confidential.
Audio Privacy & Consent Implications
Many recording and transcription laws – such as two-party consent laws in certain US states, or GDPR in Europe – are specifically concerned with the recording and storage of audio. With audio saving turned off (the default), Hedy:
Does not record or save audio files
Only holds audio data in volatile memory (RAM) for the seconds needed to process it
Does not transmit audio to any server or third party
This means the consent requirements that apply to Hedy may differ from those that apply to tools that record and store audio. The text transcript that Hedy generates stays on your device unless you explicitly choose to share or export it.
If you enable audio saving, the conversation audio will be stored locally on your device. In that case, the same consent considerations that apply to any audio recording tool may apply.
Privacy and recording laws vary by jurisdiction. We encourage users to familiarize themselves with local regulations. This information is provided for transparency, not as legal advice.
Real-Time Transcription: The Technical Challenge
Creating accurate, instantaneous transcripts on a mobile device is incredibly complex. Hedy must balance three critical factors:
Processing speed
Computational efficiency
Speech recognition accuracy
Our approach prioritizes giving you immediate insights. This means the transcript you see is a living, evolving document – capturing the essence of your conversation as it happens.
Understanding Transcript Characteristics
Transcripts generated by Hedy are unique. They're not meant to be perfect court-level recordings, but dynamic, AI-powered conversation captures. You might notice:
Occasional punctuation variations
Slight word choice differences
Informal formatting that captures conversational flow
Think of Hedy's transcript like a skilled note-taker who captures the spirit of the conversation, not a stenographic machine.
Language and Accuracy
We're committed to improving transcription across languages. Currently, English offers the most refined experience, but we're continuously expanding and refining our local processing capabilities. As phone processors become more powerful, our transcription will become increasingly sophisticated.
Improving Your Transcript Quality
While Hedy works hard to capture your conversations, you can help:
Speak clearly and at a moderate pace
Minimize background noise
Position your device optimally
Background noise, multiple overlapping speakers, and complex audio environments can challenge even the most advanced speech recognition.
Accessing and Using Your Transcripts
Hedy makes your transcripts easily accessible:
View in real-time during sessions
Review past sessions
Share or export for further analysis
Integrated with AI insights and highlights
The Future of Local Transcription
We're just getting started. Our roadmap includes:
Expanding language support
Improving processing algorithms
Enhancing real-time accuracy
Reducing computational overhead