Feature Request: Add Speech Input Support (--listen mode)

[VJLIVE]

Description:

Enable users to speak natural language commands using their microphone, which PromptShell will convert to text and then process as a query. This makes the tool more accessible, hands-free, and faster to use in some workflows.

Expected Behavior:

• Add a CLI flag:

--listen

• PromptShell will:
  • Activate the microphone
  • Convert speech to text (via offline/online models)
  • Show transcribed text as a normal query
  • Send it to the LLM
  • Display the result

Dependencies:

• Python library: speech_recognition

Test case example:

$ promptshell --listen
🎤 Listening... (press Ctrl+C to stop)

User said: "List all Python files modified in the last 2 days"

Response:
find . -name "*.py" -mtime -2

Challenges

1. Cross-platform microphone access
2. Error handling if mic access fails
3. Optional fallback to text if no speech detected

Benefits:

1. Greatly improves accessibility
2. Useful for hands-free use cases (e.g., DevOps, on-the-go scripting)
3. Helps differently-abled users interact with the shell via voice

[VJLIVE]

@Kirti-Rathi Please assign me the issue, I think this is a great feature for the project..

[Kirti-Rathi]

Assigned!

But implement --listen as a REPL command just how --config and --help works right now, rather than CLI flag

[VJLIVE]

Hello @Kirti-Rathi Sry that I was working with other projects. Yeah so, the thing is there should be a header file - portAudio.h in linux, which is not generally available unless upgraded.
Can you please confirm this?