[Feature Request] Add real-time streaming transcription as a selectable voice input mode

Open
  • Is something wrong?
#766 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
35/100
Issue type
Feature
Clarity
Mostly clear
Activity status
Active
Tech stack
android, kotlin
Domain
mobile

Research direction

Start by reading the Voice Plugin documentation mentioned in the issue and tracing the existing offline record-then-transcribe flow. Identify where voice input modes and recording controls are handled; the issue does not name specific files or tests. Done means users can select streaming or batch transcription, with silence timeout and manual controls supported in both modes.

Written by the indexing model from the issue text.

Description

enhancement

Is your feature request related to a problem?
Currently, LeanType's offline voice input operates in a record-then-transcribe workflow. The user must finish recording and press the checkmark before any recognized text is inserted into the active text field.

There does not appear to be an option for real-time streaming transcription, despite the Voice Plugin documentation describing recognized text as streaming directly into the active input field.

Describe the solution you'd like
Provide an option to choose between two voice input modes:

  1. Streaming transcription: Recognized text is progressively inserted into the active text field while the user speaks, without requiring the recording to finish first.
  2. Batch transcription: Retain the existing workflow, where audio is recorded first and transcribed when recording ends.

Ideally, both modes would support the existing silence timeout and manual recording controls.

Use case
For everyday voice typing, particularly longer messages, emails, and notes, streaming transcription would allow users to see what has been recognized as they speak.

This provides immediate feedback, makes it easier to identify recognition errors, and avoids having to wait until the end of a recording to discover that portions of speech were missed.

Batch transcription would remain useful when users prefer to dictate an entire passage before reviewing the result.

Alternatives considered
The current batch transcription workflow works for shorter dictation, but requires recording to finish before the user can review any recognized text.

Dictating in shorter segments provides more frequent feedback, but interrupts the natural flow of speaking.

Other keyboards and voice input applications offer progressive transcription.

App version & Flavor

  • LeanType: v4.3.1, Offline
  • Voice Plugin: v1.0.1
  • Whisper Base (~57 MB)
Dominant language
Kotlin
Stars
1k
Forks
59
Avg merge
14h 9m
Merged PRs (30d)
56

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from LeanBitLab/LeanType

All issues in LeanBitLab/LeanType

Similar issues

More Kotlin issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.