@tacone/prosey 0.2.6 → 0.3.0

package/README.md

@@ -1,19 +1,36 @@
+<p align="center">
+  <img src="assets/prosey-logo.svg" alt="Prosey logo" width="400"/>
+  <br>
+  <br>
+</p>
+
 # Prosey
 
-**Prosey** downloads YouTube transcripts to be printed or saved to a file.
+**Prosey** is a CLI tool to download and display video transcriptions from YouTube, so you can get the information you're looking for without wasting hours.
+
+The main use cases are:
+
+- **Summarize** a YouTube video and display it in the CLI with a single command (quicker).
+- **Read** video captions, formatted and organized in chapters like the original video (often needs more time).
+
+Both uses need an AI assistant on your system (default: [opencode](https://opencode.ai/), use `prosey config` to configure your own).
+
+If we detect a syntax highlighter on your system, you'll get a nice colored output.
 
-You can read, skim, search, copy, and manipulate the text using the tools you love the most.
+Prosey can also output or store the native json format verbatim.
 
 ## Features
 
-| Feature                              | Description                                                                                     |
-| ------------------------------------ | ----------------------------------------------------------------------------------------------- |
-| **📋&nbsp;Transcript&nbsp;download** | Fetch YouTube transcripts as plain text or JSON                                                 |
-| **🤖&nbsp;AI&nbsp;summarization**    | Pipe transcripts to any AI agent (opencode, Claude, Copilot, Codex)                             |
-| **🎨&nbsp;Smart&nbsp;pager**         | Automatic syntax highlighting if bat, glow or mdcat are installed (or configure your own pager) |
-| **🛠️&nbsp;Customization**            | Use `prosey config` to access and edit your configuration                                       |
-| **💾&nbsp;Disk&nbsp;caching**        | Transcripts and summaries cached for instant repeat                                             |
-| **📄&nbsp;Supported&nbsp;formats**   | Plain text and JSON                                                                             |
+| Feature                                    | Description                                                                                     |
+| ------------------------------------------ | ----------------------------------------------------------------------------------------------- |
+| **💡&nbsp;&nbsp;Summarize**                | Fetch and summarize a video transcription                                                       |
+| **📖&nbsp;&nbsp;Read**                     | Fetch and display the transcript formatted in Markdown                                          |
+| **📋&nbsp;&nbsp;Transcript&nbsp;download** | Fetch YouTube transcripts as plain text or JSON verbatim                                        |
+| **🤖&nbsp;&nbsp;AI&nbsp;summarization**    | Pipe transcripts to any AI agent (opencode, Claude, Copilot, Codex)                             |
+| **🎨&nbsp;&nbsp;Smart&nbsp;pager**         | Automatic syntax highlighting if bat, glow or mdcat are installed (or configure your own pager) |
+| **🛠️&nbsp;&nbsp;Customization**            | Use `prosey config` to access and edit your configuration / customize the prompts               |
+| **💾&nbsp;&nbsp;Disk&nbsp;caching**        | Transcripts and summaries are cached on disk for instant re-read                                |
+| **📄&nbsp;&nbsp;Supported&nbsp;formats**   | Plain text and JSON                                                                             |
 
 ## Quickstart
 
@@ -64,50 +81,59 @@ Grab a compiled binary from the `dist/` directory (requires no runtime).
 
 ```
 prosey [options] <video-url-or-id>
+prosey read [options] <video-url-or-id>
 prosey info [options] <video-url-or-id>
 prosey summarize [options] <video-url-or-id>
 prosey config
+prosey help
 ```
 
-Pass a full YouTube URL or a bare video ID. The transcript is printed to
-stdout by default, with video details prepended.
+`summarize` is the default command, so `prosey <id>` runs the AI summarizer.
+Use `prosey read <id>` to download and print a richly formatted transcript.
 
 The `summarize` command fetches a transcript, prepends the prompt from the
-`[summarize]` config section, and pipes the result to the configured command.
+`[summarize]` config section, and pipes the result to the configured AI command.
+
+The `read` command downloads the transcript and prints it to stdout (plain text,
+markdown, or JSON). This is useful when you don't need AI processing.
+
+The `info` command shows video metadata (title, channel, duration, views).
 
 The `config` command opens your config file in `$EDITOR` for editing. If
 `$EDITOR` is not set, the config file path is printed.
 
+The `help` command shows the help message, same as `--help`.
+
 ### Examples
 
 ```bash
-# Basic — plain text with details
+# Run the AI summarizer (default command)
 prosey 771PQEDeRmw
 
+# Download a plain transcript with details
+prosey read 771PQEDeRmw
+
 # Specify language
-prosey https://youtu.be/771PQEDeRmw --lang es
+prosey read https://youtu.be/771PQEDeRmw --lang es
 
 # Include timestamps
-prosey 771PQEDeRmw -t
+prosey read 771PQEDeRmw -t
 
-# Save to file
-prosey 771PQEDeRmw -o transcript.txt
+# Save transcript to file
+prosey read 771PQEDeRmw -o transcript.txt
 
 # JSON output (timestamps always included)
-prosey 771PQEDeRmw --json
+prosey read 771PQEDeRmw --json
 
 # Transcript only, no video details
-prosey 771PQEDeRmw --no-details
+prosey read 771PQEDeRmw --no-details
 
 # List available transcript languages
-prosey 771PQEDeRmw --list
+prosey read 771PQEDeRmw --list
 
 # Show video metadata
 prosey info 771PQEDeRmw
 
-# Summarize via the configured command
-prosey summarize 771PQEDeRmw
-
 # Edit config in $EDITOR
 prosey config
 ```
@@ -121,7 +147,7 @@ prosey config
 | `--list`                | List available transcript languages for the video, then exit.                                                  |
 | `-o`, `--output <path>` | Write output to file instead of stdout.                                                                        |
 | `--json`                | Output transcript as a JSON array. Each item includes `text`, `offset` (seconds), `duration`, and `timestamp`. |
-| `--text`                | Output as plain text (default).                                                                                |
+| `--text`                | Output as plain text (default for `prosey read`).                                                              |
 | `--details`             | Prepend video details (title, channel, duration, views, description) to the transcript (default).              |
 | `--no-details`          | Suppress video details, transcript only.                                                                       |
 | `--no-decode-entities`  | Preserve raw HTML entities (e.g. `&#39;`). Decoded by default in text mode.                                    |

package/bin/prosey

@@ -101612,6 +101612,11 @@ pager = "auto"
 # Can also be set via PROSEY_HINTS env var (yes, no, 1, 0, true, false).
 hints = true
 
+[ai]
+# Default command for AI operations (summarize, transcribe).
+# Can be overridden per-section via the command key below.
+command = "opencode run"
+
 [summarize]
 # Prompt sent to the command via stdin.
 # Customize this to change how transcripts are summarized.
@@ -101619,23 +101624,20 @@ prompt = """
 Write a comprehensive summary of the following transcription.
 """
 
-# Command to execute with the prompt and transcript piped via stdin.
-# The transcript is appended to the prompt automatically.
-#
-# Available options:
-#
-#   opencode run                    — full access (default)
-#   opencode run --permissions read  — read-only (view files, no edits)
-#
-#   claude -p "" --print            — full access (--print for clean output)
-#   claude --permission-mode plan -p "" --print  — read-only (plan/read only)
-#
-#   copilot -sp ""                  — full access (-s = silent, -p = prompt)
-#   copilot -sp "" --deny-all-tools — read-only (no shell/write access)
-#
-#   codex --sandbox default -p ""   — full access
-#   codex --sandbox read-only -p "" — read-only
-command = "opencode run"
+# Command override for summarize. Uncomment to use a different command
+# than the one specified in [ai].
+# command = "opencode run"
+
+[transcribe]
+# Prompt sent to the command via stdin.
+# Customize this to change how transcripts are formatted as markdown.
+prompt = """
+Convert this transcript to clean, readable markdown.
+"""
+
+# Command override for transcribe. Uncomment to use a different command
+# than the one specified in [ai].
+# command = "opencode run"
 `;
 async function readDefaultConfig() {
   const paths = [
@@ -101685,7 +101687,7 @@ async function resetConfig() {
 
 // src/summarize.ts
 import { spawn } from "node:child_process";
-function executeCommand(command, input, cwd) {
+var defaultExecuteCommand = (command, input, cwd) => {
   return new Promise((resolve, reject) => {
     const proc = spawn(command, [], { shell: true, stdio: "pipe", cwd });
     let stdout = "";
@@ -101706,13 +101708,16 @@ function executeCommand(command, input, cwd) {
     proc.stdin.write(input);
     proc.stdin.end();
   });
-}
-async function summarize(options) {
+};
+async function summarize(options, execCommand = defaultExecuteCommand) {
   const { prompt, command, transcript, cwd } = options;
+  if (!prompt) {
+    throw new Error("No prompt configured. A prompt is required in the config.");
+  }
   const fullPrompt = `${prompt}
 
 ${transcript}`;
-  const output = await executeCommand(command, fullPrompt, cwd);
+  const output = await execCommand(command, fullPrompt, cwd);
   const cleaned = output.startsWith(fullPrompt) ? output.slice(fullPrompt.length).replace(/\n+$/, "") : output.replace(/\n+$/, "");
   if (!cleaned || cleaned === transcript) {
     throw new Error("Summarization command returned no meaningful output");
@@ -101720,6 +101725,20 @@ ${transcript}`;
   return cleaned;
 }
 
+// src/config-resolve.ts
+function resolveSummarizeCmd(config) {
+  return config.summarize?.command ?? config.ai?.command ?? null;
+}
+function resolveSummarizePrompt(config) {
+  return config.summarize?.prompt ?? null;
+}
+function resolveTranscribeCmd(config) {
+  return config.transcribe?.command ?? config.ai?.command ?? config.summarize?.command ?? null;
+}
+function resolveTranscribePrompt(config) {
+  return config.transcribe?.prompt ?? config.summarize?.prompt ?? null;
+}
+
 // src/cache.ts
 import { createHash } from "node:crypto";
 import { readFile as readFile2, writeFile as writeFile2, mkdir as mkdir2 } from "node:fs/promises";
@@ -101756,10 +101775,62 @@ async function writeCache(dir, filename, data) {
     await mkdir2(dir, { recursive: true });
   await writeFile2(join2(dir, filename), data, "utf8");
 }
+
+// src/extract-chapters.ts
+var lineRegex = /^\s*[\[\(]?(?:(?:(\d{1,2}):)?(\d{1,2}):(\d{2}))[\]\)]?(?:\s*[-–—:.]\s*|\s+)(.+)$/;
+function extractChapters(description) {
+  const lines = description.split(`
+`);
+  const chapters = [];
+  for (const line of lines) {
+    const match = line.match(lineRegex);
+    if (!match)
+      continue;
+    const hours = match[1] ? parseInt(match[1], 10) : 0;
+    const minutes = parseInt(match[2], 10);
+    const seconds = parseInt(match[3], 10);
+    const title = match[4].trim();
+    if (!title)
+      continue;
+    const time = hours * 3600 + minutes * 60 + seconds;
+    chapters.push({ time, title });
+  }
+  chapters.sort((a2, b2) => a2.time - b2.time);
+  const seen = new Set;
+  return chapters.filter((c2) => {
+    if (seen.has(c2.time))
+      return false;
+    seen.add(c2.time);
+    return true;
+  });
+}
+function formatChaptersAsJson(chapters) {
+  if (chapters.length === 0)
+    return "not available";
+  const obj = {};
+  for (const ch of chapters) {
+    const h2 = Math.floor(ch.time / 3600);
+    const m2 = Math.floor(ch.time % 3600 / 60);
+    const s2 = ch.time % 60;
+    const key = h2 > 0 ? `${String(h2).padStart(2, "0")}:${String(m2).padStart(2, "0")}:${String(s2).padStart(2, "0")}` : `${String(m2).padStart(2, "0")}:${String(s2).padStart(2, "0")}`;
+    obj[key] = ch.title;
+  }
+  return JSON.stringify(obj);
+}
+function formatChaptersAsText(chapters) {
+  return chapters.map((c2) => {
+    const h2 = Math.floor(c2.time / 3600);
+    const m2 = Math.floor(c2.time % 3600 / 60);
+    const s2 = c2.time % 60;
+    const timeStr = h2 > 0 ? `${String(h2).padStart(2, "0")}:${String(m2).padStart(2, "0")}:${String(s2).padStart(2, "0")}` : `${String(m2).padStart(2, "0")}:${String(s2).padStart(2, "0")}`;
+    return `${timeStr} ${c2.title}`;
+  }).join(`
+`);
+}
 // package.json
 var package_default = {
   name: "@tacone/prosey",
-  version: "0.2.6",
+  version: "0.3.0",
   description: "Download YouTube video transcripts from the CLI",
   module: "src/index.ts",
   type: "module",
@@ -120385,6 +120456,10 @@ var debugApis = {
 };
 
 // src/index.ts
+process.stdout.on("error", (err) => {
+  if (err.code === "EPIPE")
+    process.exit(0);
+});
 var NAME2 = "prosey";
 var VERSION2 = package_default.version;
 var latestVersion = null;
@@ -120401,16 +120476,20 @@ function help() {
   return `${NAME2} v${VERSION2}
 
 Usage: ${NAME2} [options] <video-url-or-id>
+       ${NAME2} read [options] <video-url-or-id>
        ${NAME2} info [options] <video-url-or-id>
        ${NAME2} summarize [options] <video-url-or-id>
        ${NAME2} config
+       ${NAME2} help
 
 Download a YouTube video transcript or show video details.
 
 Commands:
+  summarize             Pipe transcript to the AI command (default command)
+  read                  Download and print a richly formatted transcript
   info                  Show video metadata (title, channel, duration, etc.)
-  summarize             Pipe transcript to the command configured in [summarize]
   config                Open config file in $EDITOR
+  help                  Show this help message
 
 Arguments:
   video-url-or-id        YouTube URL (full or short) or bare video ID
@@ -120420,14 +120499,18 @@ Options:
   -t, --timestamps       Include timestamps [MM:SS] in output.
   --list                 List available transcript languages and exit.
   -o, --output <path>    Write output to file instead of stdout.
-  --json                 Output as JSON (suppresses details).
-  --text                 Output as plain text (default).
+  --format <type>        Output format: markdown (default), text, or json.
+  --json                 Shortcut for --format json.
+  --text                 Shortcut for --format text.
+  --markdown             Shortcut for --format markdown.
   --details              Prepend video details to transcript (default, text only).
   --no-details           Suppress video details, transcript only.
   --no-decode-entities   Preserve HTML entities (decoded by default).
   --reset-config         Reset config file to defaults and exit.
   --no-cache             Skip cache and overwrite cache files.
   --no-format            Skip prettier formatting.
+  --dry-run              Print what would be sent to the AI command and exit.
+  --extract-timestamps   Extract chapter timestamps from video description.
   --no-pager             Disable pager for stdout output.
   --pager                Use pager for stdout output (default).
   --no-hints             Disable hints.
@@ -120537,7 +120620,7 @@ async function outputText(text) {
 }
 var pagerCmd = null;
 var args = process.argv.slice(2);
-if (args.length === 0 || args.includes("--help")) {
+if (args.length === 0 || args.includes("--help") || args.includes("help")) {
   console.log(help());
   exitProcess(0);
 }
@@ -120551,8 +120634,8 @@ if (args.includes("--reset-config")) {
   exitProcess(0);
 }
 var config = await loadConfig().catch(() => ({}));
-var mode = "transcript";
-var subcmdIndex = args.findIndex((a5) => a5 === "info" || a5 === "summarize" || a5 === "config");
+var mode = "summarize";
+var subcmdIndex = args.findIndex((a5) => a5 === "info" || a5 === "summarize" || a5 === "config" || a5 === "read");
 if (subcmdIndex !== -1) {
   mode = args[subcmdIndex];
   args.splice(subcmdIndex, 1);
@@ -120563,12 +120646,15 @@ var timestamps = false;
 var listOnly = false;
 var outputPath;
 var outputJson = false;
+var format3 = "markdown";
 var noDecode = false;
 var showDetails = true;
 var noCache = false;
 var noFormat = false;
 var usePager = true;
 var useHints = true;
+var dryRun = false;
+var extractTimestamps = false;
 var logLevel = "normal";
 for (let i = 0;i < args.length; i++) {
   const arg = args[i];
@@ -120592,8 +120678,27 @@ for (let i = 0;i < args.length; i++) {
     }
   } else if (arg === "--json") {
     outputJson = true;
+    format3 = "json";
   } else if (arg === "--text") {
     outputJson = false;
+    format3 = "text";
+  } else if (arg === "--markdown") {
+    format3 = "markdown";
+  } else if (arg === "--format") {
+    const val = args[++i];
+    if (val === "json") {
+      format3 = "json";
+      outputJson = true;
+    } else if (val === "text") {
+      format3 = "text";
+      outputJson = false;
+    } else if (val === "markdown") {
+      format3 = "markdown";
+      outputJson = false;
+    } else {
+      console.error("Error: --format must be text, json, or markdown");
+      exitProcess(1);
+    }
   } else if (arg === "--details") {
     showDetails = true;
   } else if (arg === "--no-details") {
@@ -120616,6 +120721,10 @@ for (let i = 0;i < args.length; i++) {
     logLevel = "verbose";
   } else if (arg === "--no-decode-entities") {
     noDecode = true;
+  } else if (arg === "--dry-run") {
+    dryRun = true;
+  } else if (arg === "--extract-timestamps") {
+    extractTimestamps = true;
   } else if (arg.startsWith("-")) {
     console.error(`Unknown option: ${arg}`);
     exitProcess(1);
@@ -120672,6 +120781,21 @@ debug("Mode:", mode);
 if (lang)
   debug("Language:", lang);
 await Promise.race([versionCheck, new Promise((r5) => setTimeout(r5, 1000))]);
+if (extractTimestamps) {
+  startTimer();
+  info("Fetching transcript...");
+  const result = await fetchTranscript(videoId, {
+    videoDetails: true,
+    lang
+  });
+  info("Transcript fetched");
+  const chapters = extractChapters(result.videoDetails.description);
+  const output = outputJson ? JSON.stringify(chapters, null, 2) + `
+` : formatChaptersAsText(chapters) + `
+`;
+  await outputText(output);
+  exitProcess(0);
+}
 try {
   if (mode === "info") {
     const result = await fetchTranscript(videoId, { videoDetails: true, lang });
@@ -120683,8 +120807,9 @@ try {
     exitProcess(0);
   }
   if (mode === "summarize") {
-    if (!config.summarize?.command) {
-      console.error("Error: [summarize] section with a command is required in config");
+    const sumCmd = resolveSummarizeCmd(config);
+    if (!sumCmd) {
+      console.error("Error: no command configured for summarize. Set [ai].command or [summarize].command in config.");
       exitProcess(1);
     }
     const cacheOpts2 = { lang, mode: "summarize", noDecode };
@@ -120713,13 +120838,24 @@ try {
       await writeCache(dir2, "transcript.json", JSON.stringify(segments2));
       debug("Cache written: transcript.json");
     }
-    const prompt = config.summarize.prompt ?? "";
+    const prompt2 = resolveSummarizePrompt(config) ?? "";
+    if (!prompt2) {
+      console.error("Error: no prompt configured. Set a prompt in the [summarize] section of your config.");
+      exitProcess(1);
+    }
     const transcriptText = toText(segments2, !noDecode);
+    if (dryRun) {
+      await outputText(`${prompt2}
+
+${transcriptText}
+`);
+      exitProcess(0);
+    }
     if (!summary) {
       info(`Summarizing...`);
       summary = await summarize({
-        prompt,
-        command: config.summarize.command,
+        prompt: prompt2,
+        command: sumCmd,
         transcript: transcriptText,
         cwd: dir2
       });
@@ -120736,6 +120872,140 @@ try {
     printLanguages(languages2);
     exitProcess(0);
   }
+  if (format3 === "markdown") {
+    const transcribeCmd = resolveTranscribeCmd(config);
+    if (!transcribeCmd) {
+      console.error("Error: no command configured for transcribe. Set [transcribe].command, [ai].command, or [summarize].command in config.");
+      exitProcess(1);
+    }
+    const cacheOpts2 = { lang, mode: "transcribe", noDecode };
+    const dir2 = cacheDir(videoId, cacheOpts2);
+    let segments2 = null;
+    let md = null;
+    startTimer();
+    let cachedInfo = null;
+    if (!noCache) {
+      const cachedSegments = await readCache(dir2, "transcript.json");
+      const cachedMd = await readCache(dir2, "transcript.md");
+      cachedInfo = await readCache(dir2, "info.json");
+      if (cachedSegments && cachedMd) {
+        info("Transcript cached");
+        debug("Cache hit:", dir2);
+        segments2 = JSON.parse(cachedSegments);
+        md = cachedMd;
+      } else {
+        debug("Cache miss:", dir2);
+      }
+    } else {
+      debug("Cache skipped (--no-cache)");
+    }
+    const prompt2 = resolveTranscribePrompt(config) ?? "";
+    if (!prompt2) {
+      console.error("Error: no prompt configured. Set a prompt in the [transcribe] or [summarize] section of your config.");
+      exitProcess(1);
+    }
+    if (!segments2) {
+      info("Fetching transcript...");
+      const opts = lang ? { lang, videoDetails: true } : { videoDetails: true };
+      const result = await fetchTranscript(videoId, opts);
+      segments2 = result.segments;
+      const infoJson = JSON.stringify({
+        title: result.videoDetails.title,
+        channel: result.videoDetails.author,
+        description: result.videoDetails.description
+      });
+      cachedInfo = infoJson;
+      const chapterValue = formatChaptersAsJson(extractChapters(result.videoDetails.description));
+      const truncatedInfo = JSON.stringify({
+        title: result.videoDetails.title,
+        channel: result.videoDetails.author,
+        description: result.videoDetails.description.slice(0, 1000)
+      });
+      const transcriptText = toText(segments2, !noDecode);
+      const structuredContent = `INFO:
+${truncatedInfo}
+
+TIMESTAMPS:
+${chapterValue}
+
+TEXT:
+${transcriptText}`;
+      if (dryRun) {
+        await outputText(`${prompt2}
+
+${structuredContent}
+`);
+        exitProcess(0);
+      }
+      info(`Transcript: ${segments2.length} segments`);
+      await writeCache(dir2, "transcript.json", JSON.stringify(segments2));
+      await writeCache(dir2, "info.json", infoJson);
+      await writeCache(dir2, "chapters.json", chapterValue);
+      debug("Cache written: transcript.json, info.json, chapters.json");
+      if (!md) {
+        info(`Transcribing...`);
+        md = await summarize({
+          prompt: prompt2,
+          command: transcribeCmd,
+          transcript: structuredContent,
+          cwd: dir2
+        });
+        info("Transcription ready");
+        await writeCache(dir2, "transcript.md", md);
+        debug("Cache written: transcript.md");
+      }
+    } else {
+      let chapterValue;
+      if (!cachedInfo) {
+        debug("Cache missing info.json, re-fetching video details");
+        const fallbackOpts = lang ? { lang, videoDetails: true } : { videoDetails: true };
+        const fallbackResult = await fetchTranscript(videoId, fallbackOpts);
+        cachedInfo = JSON.stringify({
+          title: fallbackResult.videoDetails.title,
+          channel: fallbackResult.videoDetails.author,
+          description: fallbackResult.videoDetails.description
+        });
+        await writeCache(dir2, "info.json", cachedInfo);
+        chapterValue = formatChaptersAsJson(extractChapters(fallbackResult.videoDetails.description));
+        await writeCache(dir2, "chapters.json", chapterValue);
+        debug("Cache written: info.json, chapters.json");
+      } else {
+        const cachedChapters = await readCache(dir2, "chapters.json");
+        chapterValue = cachedChapters ?? "not available";
+      }
+      const transcriptText = toText(segments2, !noDecode);
+      const cachedInfoObj = JSON.parse(cachedInfo);
+      const truncatedInfo = JSON.stringify({
+        title: cachedInfoObj.title,
+        channel: cachedInfoObj.channel,
+        description: cachedInfoObj.description.slice(0, 1000)
+      });
+      const structuredContent = `INFO:
+${truncatedInfo}
+
+TIMESTAMPS:
+${chapterValue}
+
+TEXT:
+${transcriptText}`;
+      if (!md) {
+        info(`Transcribing...`);
+        md = await summarize({
+          prompt: prompt2,
+          command: transcribeCmd,
+          transcript: structuredContent,
+          cwd: dir2
+        });
+        info("Transcription ready");
+        await writeCache(dir2, "transcript.md", md);
+        debug("Cache written: transcript.md");
+      }
+    }
+    const formatted = noFormat ? md : await formatMd(md);
+    await outputText(formatted + `
+`);
+    exitProcess(0);
+  }
   const decode = !noDecode;
   const cacheOpts = { lang, timestamps, json: outputJson, noDecode };
   const dir = cacheDir(videoId, cacheOpts);
@@ -120754,6 +121024,11 @@ try {
   } else {
     debug("Cache skipped (--no-cache)");
   }
+  const prompt = resolveTranscribePrompt(config) ?? "";
+  if (!prompt) {
+    console.error("Error: no prompt configured. Set a prompt in the [transcribe] or [summarize] section of your config.");
+    exitProcess(1);
+  }
   if (!segments) {
     info("Fetching transcript...");
     if (showDetails && !outputJson) {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tacone/prosey",
-  "version": "0.2.6",
+  "version": "0.3.0",
   "description": "Download YouTube video transcripts from the CLI",
   "module": "src/index.ts",
   "type": "module",