Update whisper-dictation extension

extensions/whisper-dictation/CHANGELOG.md

@@ -1,5 +1,13 @@
  # Changelog
 
+## [0.1.1] - 2025-06-26
+
+### Added
+- Option to both copy and paste transcibed text automatically
+- Added seperate commands for dictation and dictation with AI refinement
+  - This gives more flexibility and how and when each command is called
+- Added shortcut to skip refinement for a sesssion during the prompt selection menu (if configured)
+
  ## [0.1.0] - 2025-06-05
 
  ### Added
@@ -8,4 +16,4 @@
   - Download and manage Whisper models within Raycast
   - AI-based refinement via Raycast AI or Ollama/OpenAI-compatible APIs
   - Dictation history with browse, copy, and paste capabilities
-  - Configurable default actions (paste, copy, or manual)
+  - Configurable default actions (paste, copy, or manual)

extensions/whisper-dictation/README.md

@@ -1,31 +1,17 @@
 # 🎤 Whisper Dictation for Raycast
 
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
+
 Effortlessly convert your speech to text directly within Raycast using the power of [`whisper.cpp`](https://github.com/ggerganov/whisper.cpp). This extension provides a simple interface to record audio, transcribe and refine it locally, privately on your machine. Refine the text with custom prompts privately using ollama, or additionally with Raycast AI or any v1 (OpenAI) compatible API.
 ## ✨ Features
 
 *   **Local Transcription:** Uses `whisper.cpp` running locally on your machine through Raycast.
-*   **Refine with AI:** Optionally refine the transcribed text using Raycast AI or any OpenAI (v1) compatible API, such as Anthropic, OpenAI, a local Ollama server.
-*   **Download Models:** Download a variety of whisper models from right within the extension.
-*   **Dictation History:** Saves all transcriptions locally with timestamps which can be browsed, copied and pasted using the Dictation History command.
+*   **Refine with AI** Optionally refine the transcribed text using Raycast AI or any OpenAI (v1) compatible API, such as Anthropic, OpenAI, a local Ollama server.
+    **Download Models** Download a variety of whisper models from right within the extension.
+    **Dictation History** Saves all transcriptions locally with timestamps which can be browsed, copied and pasted using the Dictation History command.
 *   **Simple Interface:** Start recording, press Enter to stop, copy or paste directly into your active window.
 *   **Configurable Output:** Choose to choose, or automatically paste/copy to clipboard. 
 
-## 📚 Table of Contents
-
-*   [Features](#-features)
-*   [Requirements](#-requirements)
-*   [Installation](#-installation)
-    *   [1. Prerequisites](#1-prerequisites)
-    *   [2. Install the Extension](#2-install-the-extension)
-*   [Configuration](#️-configuration)
-*   [Usage](#-usage)
-*   [Refine with AI](#-refine-with-ai)
-*   [Models](#-models)
-*   [Troubleshooting](#-troubleshooting)
-*   [Contributing](#-contributing)
-*   [License](#-license)
-*   [Acknowledgements](#-acknowledgements)
-
 ## ⚠️ Requirements
 
 Before installing the extension, you need the following installed and configured on your system:
@@ -41,6 +27,10 @@ Before installing the extension, you need the following installed and configured
     *   The easiest way to install it on macOS is with [Homebrew](https://brew.sh/): `brew install sox`
     *The extension currently default for `sox` to be at `/opt/homebrew/bin/sox`. If yours is installed somewhere else, point the extension to it's executable in preferences.
 
+## 🚀 Installation
+
+**This ectension is now available to download from the [Raycast Store](https://www.raycast.com/finjo/whisper-dictation). However if you'd prefer to build from source see below**
+
 ## ⚙️ Configuration
 
 After installing, you have to configure the extension preferences in Raycast, if you installed both SoX and whisper-cpp using homebrew, and download a model using the extension this should all be pre-configured for you, the extension will also confirm both SoX and whisper-cli path on first launch which will allow you to immediately start using simple dictation once configured:
@@ -161,4 +151,14 @@ The extension downloader currently supports the following whisper models, howeve
             *   **External API:** https://api.anthropic.com
     *   If using Raycast AI make sure that you have paid for Raycast Pro
 
----
+## 📄 License
+
+This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details (or state MIT directly if no file exists).
+
+## ❤️ Acknowledgements
+
+*   [Georgi Gerganov](https://github.com/ggerganov) for the amazing [`whisper.cpp`](https://github.com/ggerganov/whisper.cpp) project.
+*   The [Raycast](https://raycast.com/) team for the fantastic platform and API.
+*   [SoX - Sound eXchange](https://github.com/chirlu/sox) developers.
+
+---

extensions/whisper-dictation/package.json

@@ -19,11 +19,18 @@
     "transcribed.png"
   ],
   "commands": [
+    {
+      "name": "dictate-simple",
+      "title": "Dictate",
+      "subtitle": "Whisper Dictation",
+      "description": "Convert speech to text using Whisper (no AI refinement)",
+      "mode": "view"
+    },
     {
       "name": "dictate",
-      "title": "Dictate Text",
+      "title": "Dictate with AI",
       "subtitle": "Whisper Dictation",
-      "description": "Convert speech to text using Whisper",
+      "description": "Convert speech to text with AI refinement options",
       "mode": "view"
     },
     {
@@ -84,16 +91,20 @@
       "default": "none",
       "data": [
         {
-          "title": "Paste Text",
-          "value": "paste"
+          "title": "None (Show Options)",
+          "value": "none"
         },
         {
           "title": "Copy to Clipboard",
           "value": "copy"
         },
         {
-          "title": "None (Show Options)",
-          "value": "none"
+          "title": "Paste Text",
+          "value": "paste"
+        },
+        {
+          "title": "Copy & Paste Text",
+          "value": "copy_paste"
         }
       ]
     },

extensions/whisper-dictation/src/dictate-simple.tsx

@@ -0,0 +1,436 @@
+import {
+  ActionPanel,
+  Form,
+  Action,
+  showToast,
+  Toast,
+  closeMainWindow,
+  Icon,
+  Detail,
+  getPreferenceValues,
+  environment,
+  LocalStorage,
+  launchCommand,
+  LaunchType,
+  showHUD,
+  openExtensionPreferences,
+  PopToRootType,
+} from "@raycast/api";
+import { useState, useEffect, useRef, useCallback } from "react";
+import type { ChildProcessWithoutNullStreams } from "child_process";
+import path from "path";
+import fs from "fs";
+import crypto from "crypto";
+import { useConfiguration } from "./hooks/useConfiguration";
+import { useRecording } from "./hooks/useRecording";
+import { useTranscription } from "./hooks/useTranscription";
+
+interface TranscriptionHistoryItem {
+  id: string;
+  timestamp: number;
+  text: string;
+}
+
+const AUDIO_FILE_PATH = path.join(environment.supportPath, "raycast_dictate_audio.wav");
+const HISTORY_STORAGE_KEY = "dictationHistory";
+
+type CommandState =
+  | "configuring"
+  | "configured_waiting_selection"
+  | "selectingPrompt"
+  | "idle"
+  | "recording"
+  | "transcribing"
+  | "done"
+  | "error";
+
+interface Config {
+  execPath: string;
+  modelPath: string;
+  soxPath: string;
+}
+
+export default function SimpleDictateCommand() {
+  const [state, setState] = useState<CommandState>("configuring");
+  const [transcribedText, setTranscribedText] = useState<string>("");
+  const [errorMessage, setErrorMessage] = useState<string>("");
+  const soxProcessRef = useRef<ChildProcessWithoutNullStreams | null>(null);
+  const [waveformSeed, setWaveformSeed] = useState<number>(0);
+  const [config, setConfig] = useState<Config | null>(null);
+
+  const preferences = getPreferenceValues<Preferences>();
+  const DEFAULT_ACTION = preferences.defaultAction || "none";
+
+  const cleanupAudioFile = useCallback(() => {
+    fs.promises
+      .unlink(AUDIO_FILE_PATH)
+      .then(() => console.log("Cleaned up audio file."))
+      .catch((err) => {
+        if (err.code !== "ENOENT") {
+          console.error("Error cleaning up audio file:", err.message);
+        }
+      });
+  }, []);
+
+  useConfiguration(setState, setConfig, setErrorMessage);
+
+  const saveTranscriptionToHistory = useCallback(async (text: string) => {
+    if (!text || text === "[BLANK_AUDIO]") return;
+
+    try {
+      const newItem: TranscriptionHistoryItem = {
+        id: crypto.randomUUID(),
+        timestamp: Date.now(),
+        text: text,
+      };
+
+      let history: TranscriptionHistoryItem[] = [];
+      const existingHistory = await LocalStorage.getItem<string>(HISTORY_STORAGE_KEY);
+      if (existingHistory) {
+        try {
+          history = JSON.parse(existingHistory);
+        } catch (parseError) {
+          console.error("Failed to parse history from LocalStorage:", parseError);
+          await showToast({
+            style: Toast.Style.Failure,
+            title: "Warning",
+            message: "Could not read previous dictation history. Clearing history.",
+          });
+          history = [];
+        }
+      }
+
+      history.unshift(newItem);
+
+      const MAX_HISTORY_ITEMS = 100;
+      if (history.length > MAX_HISTORY_ITEMS) {
+        history = history.slice(0, MAX_HISTORY_ITEMS);
+      }
+
+      await LocalStorage.setItem(HISTORY_STORAGE_KEY, JSON.stringify(history));
+      console.log("Saved transcription to history.");
+    } catch (error) {
+      console.error("Failed to save transcription to history:", error);
+      await showToast({
+        style: Toast.Style.Failure,
+        title: "Error",
+        message: "Failed to save transcription to history.",
+      });
+    }
+  }, []);
+
+  const { startTranscription, handlePasteAndCopy } = useTranscription({
+    config,
+    preferences,
+    setState,
+    setErrorMessage,
+    setTranscribedText,
+    refineText: async (text: string) => text, // No AI refinement
+    saveTranscriptionToHistory,
+    cleanupAudioFile,
+    aiErrorMessage: "",
+    skipAIForSession: true, // Always skip AI for simple dictate
+  });
+
+  // Function to stop recording and transcribe
+  const stopRecordingAndTranscribe = useCallback(async () => {
+    console.log(`stopRecordingAndTranscribe called. Current state: ${state}`);
+
+    if (state !== "recording") {
+      console.warn(`stopRecordingAndTranscribe: State is '${state}', expected 'recording'. Aborting.`);
+      return;
+    }
+
+    const processToStop = soxProcessRef.current;
+
+    if (processToStop) {
+      console.log(`Attempting to stop recording process PID: ${processToStop.pid}...`);
+      soxProcessRef.current = null;
+      console.log("Cleared sox process ref.");
+      try {
+        if (!processToStop.killed) {
+          process.kill(processToStop.pid!, "SIGTERM");
+          console.log(`Sent SIGTERM to PID ${processToStop.pid}`);
+          await new Promise((resolve) => setTimeout(resolve, 100));
+        } else {
+          console.log(`Process ${processToStop.pid} was already killed.`);
+        }
+      } catch (e) {
+        if (e instanceof Error && "code" in e && e.code !== "ESRCH") {
+          console.warn(`Error stopping sox process PID ${processToStop.pid}:`, e);
+        } else {
+          console.log(`Process ${processToStop.pid} likely already exited.`);
+        }
+      }
+    } else {
+      console.warn("stopRecordingAndTranscribe: No active sox process reference found to stop.");
+    }
+
+    await startTranscription();
+  }, [state, startTranscription]);
+
+  const { restartRecording } = useRecording(state, config, setState, setErrorMessage, soxProcessRef);
+
+  // Effect to automatically transition to idle after configuration (skip prompt selection)
+  useEffect(() => {
+    if (state === "configured_waiting_selection") {
+      setState("idle");
+    }
+  }, [state]);
+
+  useEffect(() => {
+    let intervalId: NodeJS.Timeout | null = null;
+    if (state === "recording") {
+      intervalId = setInterval(() => {
+        setWaveformSeed((prev) => prev + 1);
+      }, 150);
+    }
+    return () => {
+      if (intervalId) {
+        clearInterval(intervalId);
+      }
+    };
+  }, [state]);
+
+  useEffect(() => {
+    if (state === "transcribing") {
+      const timer = setTimeout(() => {
+        startTranscription();
+      }, 100);
+      return () => clearTimeout(timer);
+    }
+  }, [state, startTranscription]);
+
+  useEffect(() => {
+    return () => {
+      const currentProcess = soxProcessRef.current;
+      if (currentProcess && !currentProcess.killed) {
+        console.log(`Cleanup: Killing process PID: ${currentProcess.pid}`);
+        try {
+          process.kill(currentProcess.pid!, "SIGKILL");
+        } catch (e) {
+          if (e instanceof Error && "code" in e && e.code !== "ESRCH") {
+            console.warn("Cleanup: Error sending SIGKILL:", e);
+          }
+        }
+      }
+      cleanupAudioFile();
+    };
+  }, [cleanupAudioFile]);
+
+  const generateWaveformMarkdown = useCallback(() => {
+    const waveformHeight = 18;
+    const waveformWidth = 105;
+    let waveform = "```\n";
+    waveform += "RECORDING AUDIO... PRESS ENTER TO STOP\n\n";
+
+    for (let y = 0; y < waveformHeight; y++) {
+      let line = "";
+      for (let x = 0; x < waveformWidth; x++) {
+        const baseAmplitude1 = Math.sin((x / waveformWidth) * Math.PI * 4) * 0.3;
+        const baseAmplitude2 = Math.sin((x / waveformWidth) * Math.PI * 8) * 0.15;
+        const baseAmplitude3 = Math.sin((x / waveformWidth) * Math.PI * 2) * 0.25;
+        const baseAmplitude = baseAmplitude1 + baseAmplitude2 + baseAmplitude3;
+        const randomFactor = Math.sin(x + waveformSeed * 0.3) * 0.2;
+        const amplitude = baseAmplitude + randomFactor;
+        const normalizedAmplitude = (amplitude + 0.7) * waveformHeight * 0.5;
+        const distFromCenter = Math.abs(y - waveformHeight / 2);
+        const shouldDraw = distFromCenter < normalizedAmplitude;
+
+        if (shouldDraw) {
+          const intensity = 1 - distFromCenter / normalizedAmplitude;
+          if (intensity > 0.8) line += "█";
+          else if (intensity > 0.6) line += "▓";
+          else if (intensity > 0.4) line += "▒";
+          else if (intensity > 0.2) line += "░";
+          else line += "·";
+        } else {
+          line += " ";
+        }
+      }
+      waveform += line + "\n";
+    }
+    waveform += "```";
+    return waveform;
+  }, [waveformSeed]);
+
+  const getActionPanel = useCallback(() => {
+    switch (state) {
+      case "recording":
+        return (
+          <ActionPanel>
+            <Action title="Stop and Transcribe" icon={Icon.Stop} onAction={stopRecordingAndTranscribe} />
+            <Action
+              title="Cancel Recording"
+              icon={Icon.XMarkCircle}
+              shortcut={{ modifiers: ["cmd"], key: "." }}
+              onAction={() => {
+                const processToStop = soxProcessRef.current;
+                if (processToStop && !processToStop.killed) {
+                  try {
+                    process.kill(processToStop.pid!, "SIGKILL");
+                    console.log(`Cancel Recording: Sent SIGKILL to PID ${processToStop.pid}`);
+                  } catch {
+                    /* Ignore ESRCH */
+                  }
+                  soxProcessRef.current = null;
+                }
+                cleanupAudioFile();
+                closeMainWindow({ clearRootSearch: true, popToRootType: PopToRootType.Immediate });
+              }}
+            />
+            <Action
+              title="Retry Recording"
+              icon={Icon.ArrowClockwise}
+              shortcut={{ modifiers: ["cmd"], key: "r" }}
+              onAction={() => {
+                console.log("Retry Recording action triggered.");
+                cleanupAudioFile();
+                restartRecording();
+              }}
+            />
+          </ActionPanel>
+        );
+      case "done":
+        return (
+          <ActionPanel>
+            <Action.CopyToClipboard
+              title={DEFAULT_ACTION === "copy" ? "Copy Text (Default)" : "Copy Text"}
+              content={transcribedText}
+              shortcut={{ modifiers: ["cmd"], key: "enter" }}
+              onCopy={() => closeMainWindow({ clearRootSearch: true, popToRootType: PopToRootType.Immediate })}
+            />
+            <Action.Paste
+              title={DEFAULT_ACTION === "paste" ? "Paste Text (Default)" : "Paste Text"}
+              content={transcribedText}
+              shortcut={{ modifiers: ["cmd", "shift"], key: "enter" }}
+              onPaste={() => closeMainWindow({ clearRootSearch: true, popToRootType: PopToRootType.Immediate })}
+            />
+            <Action
+              title={DEFAULT_ACTION === "copy_paste" ? "Copy & Paste Text (Default)" : "Copy & Paste Text"}
+              icon={Icon.Clipboard}
+              shortcut={{ modifiers: ["cmd", "opt"], key: "enter" }}
+              onAction={() => handlePasteAndCopy(transcribedText)}
+            />
+            <Action
+              title="View History"
+              icon={Icon.List}
+              shortcut={{ modifiers: ["cmd"], key: "h" }}
+              onAction={async () => {
+                await launchCommand({ name: "dictation-history", type: LaunchType.UserInitiated });
+              }}
+            />
+            <Action title="Close" icon={Icon.XMarkCircle} onAction={closeMainWindow} />
+          </ActionPanel>
+        );
+      case "transcribing":
+        return null;
+      case "error":
+        return (
+          <ActionPanel>
+            <Action
+              title="Open Extension Preferences"
+              icon={Icon.Gear}
+              onAction={() => {
+                openExtensionPreferences();
+                closeMainWindow({ clearRootSearch: true, popToRootType: PopToRootType.Immediate });
+              }}
+            />
+            <Action
+              title="Retry (Reopen Command)"
+              icon={Icon.ArrowClockwise}
+              onAction={() => {
+                showHUD("Please reopen the Simple Dictate command.");
+                closeMainWindow({ clearRootSearch: true, popToRootType: PopToRootType.Immediate });
+              }}
+            />
+            <Action
+              title="Download Model"
+              icon={Icon.Download}
+              onAction={async () => {
+                await launchCommand({ name: "download-model", type: LaunchType.UserInitiated });
+              }}
+            />
+            <Action title="Close" icon={Icon.XMarkCircle} onAction={closeMainWindow} />
+          </ActionPanel>
+        );
+      default:
+        return (
+          <ActionPanel>
+            <Action title="Start Recording" icon={Icon.Microphone} onAction={() => setState("recording")} />
+            <Action
+              title="View History"
+              icon={Icon.List}
+              shortcut={{ modifiers: ["cmd"], key: "h" }}
+              onAction={async () => {
+                await launchCommand({ name: "dictation-history", type: LaunchType.UserInitiated });
+              }}
+            />
+            <Action
+              title="Open Extension Preferences"
+              icon={Icon.Gear}
+              onAction={() => {
+                openExtensionPreferences();
+                closeMainWindow({ clearRootSearch: true, popToRootType: PopToRootType.Immediate });
+              }}
+            />
+            <Action title="Close" icon={Icon.XMarkCircle} onAction={closeMainWindow} />
+          </ActionPanel>
+        );
+    }
+  }, [
+    state,
+    stopRecordingAndTranscribe,
+    transcribedText,
+    cleanupAudioFile,
+    DEFAULT_ACTION,
+    handlePasteAndCopy,
+    restartRecording,
+  ]);
+
+  if (state === "configuring") {
+    return <Detail isLoading={true} markdown={"Checking Whisper configuration..."} />;
+  }
+
+  if (state === "recording") {
+    const waveformMarkdown = generateWaveformMarkdown();
+    return <Detail markdown={waveformMarkdown} actions={getActionPanel()} />;
+  }
+
+  return (
+    <Form
+      isLoading={state === "transcribing"}
+      actions={getActionPanel()}
+      navigationTitle={
+        state === "transcribing"
+          ? "Transcribing..."
+          : state === "done"
+            ? "Transcription Result"
+            : state === "error"
+              ? "Error"
+              : "Simple Dictation"
+      }
+    >
+      {state === "error" && <Form.Description title="Error" text={errorMessage} />}
+      {(state === "done" || state === "transcribing" || state === "idle") && (
+        <Form.TextArea
+          id="dictatedText"
+          title={state === "done" ? "Dictated Text" : ""}
+          placeholder={
+            state === "transcribing"
+              ? "Transcribing audio..."
+              : state === "done"
+                ? "Transcription result"
+                : "Press Enter to start recording..."
+          }
+          value={state === "done" ? transcribedText : ""}
+          onChange={setTranscribedText}
+        />
+      )}
+      {state === "transcribing" && <Form.Description text="Processing audio, please wait..." />}
+      {state === "idle" && (
+        <Form.Description text="Simple dictation without AI refinement. Press Enter to start recording." />
+      )}
+    </Form>
+  );
+}

extensions/whisper-dictation/src/dictate.tsx

@@ -64,12 +64,13 @@ interface Config {
   soxPath: string;
 }
 
-export default function Command() {
+export default function DictateWithAICommand() {
   const [state, setState] = useState<CommandState>("configuring");
   const [transcribedText, setTranscribedText] = useState<string>("");
   const [errorMessage, setErrorMessage] = useState<string>("");
   const [aiErrorMessage, setAiErrorMessage] = useState<string>("");
   const [selectedSessionPrompt, setSelectedSessionPrompt] = useState<AIPrompt | null>(null);
+  const [skipAIForSession, setSkipAIForSession] = useState<boolean>(false);
   const soxProcessRef = useRef<ChildProcessWithoutNullStreams | null>(null);
   const [waveformSeed, setWaveformSeed] = useState<number>(0);
   const [config, setConfig] = useState<Config | null>(null);
@@ -144,6 +145,13 @@ export default function Command() {
     closeMainWindow({ clearRootSearch: true, popToRootType: PopToRootType.Immediate });
   }, [cleanupAudioFile]);
 
+  // Handle skipping AI refinement entirely for current session
+  const handleSkipAIRefinement = useCallback(() => {
+    setSkipAIForSession(true);
+    setSelectedSessionPrompt(null);
+    setState("idle");
+  }, []);
+
   // Handle skipping prompt selection, will use currently active prompt or first prompt
   const handleSkipAndUseActivePrompt = useCallback(async () => {
     try {
@@ -275,11 +283,12 @@ export default function Command() {
   useEffect(() => {
     if (preferences.aiRefinementMethod === "disabled") {
       setSelectedSessionPrompt(null);
+      setSkipAIForSession(false);
     }
   }, [preferences.aiRefinementMethod]);
 
   // Use transcription hook
-  const { startTranscription } = useTranscription({
+  const { startTranscription, handlePasteAndCopy } = useTranscription({
     config,
     preferences,
     setState,
@@ -289,6 +298,7 @@ export default function Command() {
     saveTranscriptionToHistory,
     cleanupAudioFile,
     aiErrorMessage,
+    skipAIForSession,
   });
 
   // Function to stop recording and transcribe via hook
@@ -421,6 +431,12 @@ export default function Command() {
               shortcut={{ modifiers: ["cmd", "shift"], key: "enter" }}
               onPaste={() => closeMainWindow({ clearRootSearch: true, popToRootType: PopToRootType.Immediate })} // Close after paste
             />
+            <Action
+              title={DEFAULT_ACTION === "copy_paste" ? "Copy & Paste Text (Default)" : "Copy & Paste Text"}
+              icon={Icon.Clipboard}
+              shortcut={{ modifiers: ["cmd", "opt"], key: "enter" }}
+              onAction={() => handlePasteAndCopy(transcribedText)}
+            />
             <Action
               title="View History"
               icon={Icon.List}
@@ -502,6 +518,12 @@ export default function Command() {
                   onAction={handleSkipAndUseActivePrompt}
                   shortcut={{ modifiers: ["cmd"], key: "s" }}
                 />
+                <Action
+                  title="Skip AI Refinement"
+                  icon={Icon.XMarkCircle}
+                  onAction={handleSkipAIRefinement}
+                  shortcut={{ modifiers: ["cmd", "shift"], key: "s" }}
+                />
                 <Action title="Skip & Continue" icon={Icon.ArrowRight} onAction={handlePromptSelectionCancel} />
               </ActionPanel>
             }
@@ -530,6 +552,12 @@ export default function Command() {
                       onAction={handleSkipAndUseActivePrompt}
                       shortcut={{ modifiers: ["cmd"], key: "s" }}
                     />
+                    <Action
+                      title="Skip AI Refinement"
+                      icon={Icon.XMarkCircle}
+                      onAction={handleSkipAIRefinement}
+                      shortcut={{ modifiers: ["cmd", "shift"], key: "s" }}
+                    />
                     <Action
                       title="Configure AI"
                       icon={Icon.Gear}
@@ -557,7 +585,9 @@ export default function Command() {
   if (state === "recording") {
     let refinementSection = "";
 
-    if (selectedSessionPrompt) {
+    if (skipAIForSession) {
+      refinementSection = `**AI Refinement: Skipped for this session**\n\n`;
+    } else if (selectedSessionPrompt) {
       refinementSection = `**AI Refinement: ${selectedSessionPrompt.name}**\n\n`;
     } else if (isRefinementActive && currentRefinementPrompt) {
       const activePrompt = prompts.find((p) => p.prompt === currentRefinementPrompt);

extensions/whisper-dictation/src/hooks/useTranscription.ts

@@ -44,6 +44,7 @@ interface UseTranscriptionProps {
   saveTranscriptionToHistory: (text: string) => Promise<void>;
   cleanupAudioFile: () => void;
   aiErrorMessage: string;
+  skipAIForSession: boolean;
 }
 /**
  * Hook that manages audio transcription using Whisper CLI.
@@ -68,13 +69,37 @@ export function useTranscription({
   saveTranscriptionToHistory,
   cleanupAudioFile,
   aiErrorMessage,
+  skipAIForSession,
 }: UseTranscriptionProps) {
+  const handlePasteAndCopy = useCallback(
+    async (text: string) => {
+      try {
+        await Clipboard.copy(text);
+        await Clipboard.paste(text);
+        await showHUD("Copied and pasted transcribed text");
+      } catch (error) {
+        console.error("Error during copy and paste:", error);
+        showFailureToast(error, { title: "Failed to copy and paste text" });
+      }
+      await Promise.all([
+        cleanupAudioFile(),
+        closeMainWindow({ clearRootSearch: true, popToRootType: PopToRootType.Immediate }),
+      ]);
+    },
+    [cleanupAudioFile],
+  );
+
   const handleTranscriptionResult = useCallback(
     async (rawText: string) => {
       let finalText = rawText;
 
-      // Apply AI refinement if enabled and text is not empty
-      if (preferences.aiRefinementMethod !== "disabled" && rawText && rawText !== "[BLANK_AUDIO]") {
+      // Apply AI refinement if enabled and text is not empty and not skipped for session
+      if (
+        preferences.aiRefinementMethod !== "disabled" &&
+        !skipAIForSession &&
+        rawText &&
+        rawText !== "[BLANK_AUDIO]"
+      ) {
         try {
           finalText = await refineText(rawText);
         } catch (error) {
@@ -110,10 +135,12 @@ export function useTranscription({
         await handleClipboardActionAndClose("paste", finalText);
       } else if (DEFAULT_ACTION === "copy") {
         await handleClipboardActionAndClose("copy", finalText);
+      } else if (DEFAULT_ACTION === "copy_paste") {
+        await handlePasteAndCopy(finalText);
       } else {
         // Action is "none", stay in "done" state
         // Show success toast only if AI didn't fail (or wasn't used)
-        if (preferences.aiRefinementMethod === "disabled" || !aiErrorMessage) {
+        if (preferences.aiRefinementMethod === "disabled" || skipAIForSession || !aiErrorMessage) {
           await showToast({ style: Toast.Style.Success, title: "Transcription complete" });
         }
         // Clean up file when staying in 'done' state
@@ -128,6 +155,8 @@ export function useTranscription({
       setState,
       cleanupAudioFile,
       aiErrorMessage,
+      handlePasteAndCopy,
+      skipAIForSession,
     ],
   );
 
@@ -229,5 +258,5 @@ export function useTranscription({
     );
   }, [config, setState, setErrorMessage, handleTranscriptionResult, cleanupAudioFile]);
 
-  return { startTranscription };
+  return { startTranscription, handlePasteAndCopy };
 }