claude-code - 💡(How to fix) Fix [Feature Request] Add native voice input with multilingual support and technical identifier preservation

[{"error":"McpToolCallError: Caught Exception. Error: Request failed: HTTPSConnectionPool(host='127.0.0.1', port=27124): Max retries exceeded with url: /search/simple/?query=%D0%B0%D0%B2%D1%82%D0%BE%D0%BC%D0%B0%D1%82%D0%B8%D0%B7%D0%B0%D1%86%D0%B8%D1%8F+%D0%B1%D0%B5%D0%B7+%D0%B4%D0%B5%D0%BF%D0%BB%D0%BE%D1%8F&contextLength=200 (Caused by NewConnectionError("HTTPSConnection(host='127.0.0.1', port=27124): Failed to establish a new connection: [Errno 111] Connection refused"))\n at ZX7 (/$bunfs/root/src/entrypoints/cli.js:4761:63)\n at processTicksAndRejections (native:7:39)","timestamp":"2026-04-20T13:14:16.239Z"},{"error":"McpToolCallError: Caught Exception. Error: Request failed: HTTPSConnectionPool(host='127.0.0.1', port=27124): Max retries exceeded with url: /vault/KMS/Sessions/ (Caused by NewConnectionError("HTTPSConnection(host='127.0.0.1', port=27124): Failed to establish a new connection: [Errno 111] Connection refused"))\n at ZX7 (/$bunfs/root/src/entrypoints/cli.js:4761:63)\n at processTicksAndRejections (native:7:39)","timestamp":"2026-04-20T13:14:16.265Z"}]

TL;DR

Implementing a native voice input mode with automatic language detection, smart identifier preservation, and punctuation normalization could address the current limitations of using external transcription tools with Claude Code.

Guidance

• Investigate integrating Whispering's whisper-large-v3-turbo model with the "Recommended" preset to improve Russian transcription quality.
• Develop a system to preserve technical identifiers (e.g., variable names, filenames, paths) in Latin script, regardless of the dictation language.
• Implement punctuation normalization to replace smart quotes and em-dashes with standard characters before parsing the transcribed text.
• Consider adding a "Voice → Plan → Approve → Execute" mode to ensure user confirmation before executing code generated from voice input.

Example

No specific code example can be provided without more context, but the integration might involve using the Whispering API to transcribe voice input and then processing the text to preserve identifiers and normalize punctuation.

Notes

The provided error messages suggest connection issues with the HTTPSConnectionPool, but these may not be directly related to the proposed feature. The focus should be on developing a native voice input mode that addresses the current pain points with external transcription tools.

Recommendation

Apply a workaround by integrating an existing transcription service like Whispering, while developing a native voice input mode to provide a more seamless experience for users. This approach allows for immediate improvement while working towards a more comprehensive solution.