anthropics / anthropics/claude-code

Voice dictation: truncated copy of the sent message reappears in the composer

Open
#95,564 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

area:desktop bug has repro platform:windows
Dominant language
TypeScript
Stars
147k
Forks
24k
PR merge metrics
PR metrics pending

Description

Environment

  • Claude Code desktop app (Code tab), Windows 11 Pro, build 10.0.26200
  • Input method: voice dictation via the microphone button in the composer

Description

After dictating a message with the microphone and sending it, a partial copy of that same message automatically reappears in the composer input field.

The reappearing text is truncated at the beginning — it starts mid-sentence, missing roughly the first line of what was dictated. The rest matches the message that was just sent.

This happens every time dictation is used. The leftover text has to be selected and deleted manually before the next message can be typed or dictated, otherwise it gets prepended to the next message.

Steps to reproduce

  1. Click the microphone button in the composer
  2. Dictate a message of several sentences
  3. Turn the microphone off
  4. Send the message
  5. Look at the composer input field

Expected behavior

The composer is empty after sending.

Actual behavior

The composer contains a truncated copy of the message that was just sent, missing the first portion of the text.

Workaround

Click into the composer, Ctrl+A, Delete before each new message.

Notes

Appears to be a speech-to-text buffer that is not cleared on send and gets re-inserted when the input field regains focus.

Contributor guide

No contributing guide indexed for this repository

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.

Research direction

Start with the composer’s microphone and send flow described in the reproduction steps, then inspect how dictated text is retained when the input regains focus. Reproduce on Windows 11 with voice dictation and verify that sending leaves the composer empty and does not reinsert any truncated text.

Written by the indexing model from the issue text.

Assessment

Tech stack
typescript
Domain
desktop, frontend
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
68/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.