feat(blog-to-podcast): turn a blog post into a podcast episode

README.md

@@ -13,6 +13,7 @@ Demos with a **Live** link run in your browser at [demos.speechify.ai](https://d
 | [`demos/ai-sdk-speechify-speech/`](./demos/ai-sdk-speechify-speech) | Next.js + AI SDK | [Open](https://demos.speechify.ai/ai-sdk-speechify-speech) | A custom Speechify speech model for the AI SDK's generateSpeech, wired into a Next.js page. The API key stays server-side. |
 | [`demos/multilingual-voiceover/`](./demos/multilingual-voiceover) | Next.js + simba-3.0 | [Open](https://demos.speechify.ai/multilingual-voiceover) | Generate the same line in six languages with a locale-matched simba-3.0 voice. The API key stays server-side. |
 | [`demos/captions-speech-marks/`](./demos/captions-speech-marks) | TypeScript (native) |  | Synthesizes audio and builds a WebVTT caption file from the speech marks the API returns in the same response. Karaoke-highlight HTML demo included. |
+| [`demos/blog-to-podcast/`](./demos/blog-to-podcast) | TypeScript (SDK) |  | Point it at any blog post URL, get a podcast-style MP3 out: fetches the article, extracts the text, chunks on sentence boundaries, synthesizes each chunk via the Speechify API, and stitches the parts together with ffmpeg. |
 | [`demos/voice-cloning-narration/`](./demos/voice-cloning-narration) | TypeScript (native) |  | Clones a voice from a 10-30 sec WAV sample, synthesizes with the new voice, deletes the clone. End-to-end lifecycle. |
 | [`demos/audiobook-pipeline/`](./demos/audiobook-pipeline) | Python (SDK) |  | Chunks long-form text on sentence boundaries, synthesizes each chunk via the Speechify Python SDK, concatenates the MP3s with ffmpeg. |
 | [`demos/livekit-agent-speechify-python/`](./demos/livekit-agent-speechify-python) | Python (LiveKit) |  | Real-time voice assistant using LiveKit's official livekit-plugins-speechify package for TTS, with Deepgram STT and an OpenAI LLM in a LiveKit AgentSession. |

demos/blog-to-podcast/.env.example

@@ -0,0 +1 @@
+SPEECHIFY_API_KEY=your_api_key_here

demos/blog-to-podcast/README.md

@@ -0,0 +1,37 @@
+# Blog post to podcast
+
+Point it at any blog post URL, get a podcast-style MP3 out. Pairs with the
+Speechify post *"Turn this blog post into a podcast episode with the Speechify
+API"*.
+
+## What you get
+
+A single-file Node script that turns a web page into an episode:
+
+1. fetches the URL and pulls the readable text out of the `<article>`/`<main>` HTML,
+2. chunks it on paragraph/sentence boundaries (each call stays under the API input cap),
+3. synthesizes each chunk via the Speechify API,
+4. stitches the parts into one MP3 with ffmpeg.
+
+## Run it yourself
+
+```bash
+cp .env.example .env      # paste your key into .env
+npm install
+
+npm start -- https://speechify.ai/blog/some-post -v lily --model simba-english
+# ...writes output/<host>-<ts>/part-000.mp3, part-001.mp3, ...
+
+./concat.sh output/<host>-<ts> podcast-episode.mp3
+```
+
+Prerequisites: Node 20+, ffmpeg on your PATH, a Speechify API key from
+[platform.speechify.ai/api-keys](https://platform.speechify.ai/api-keys).
+
+## Notes
+
+- The extractor is naive on purpose — it strips script/style/nav/footer blocks
+  and tags from the page. JS-rendered sites will yield nothing; point it at the
+  raw HTML or a static post instead.
+- Long posts are chunked so every synthesis call stays under the input-size
+  limit; the parts are concatenated losslessly (`-c copy`).

demos/blog-to-podcast/concat.sh

@@ -0,0 +1,16 @@
+#!/usr/bin/env bash
+# Concatenate the per-chunk MP3s from output/<dir> into one file.
+# Usage: ./concat.sh output/<dir> [output.mp3]
+set -euo pipefail
+
+src="${1:?usage: ./concat.sh <output-dir> [output.mp3]}"
+out="${2:-podcast-episode.mp3}"
+list="$(mktemp -t blog-to-podcast.XXXXXX)"
+trap 'rm -f "$list"' EXIT
+
+for f in "$src"/part-*.mp3; do
+  printf "file '%s'\n" "$(cd "$(dirname "$f")" && pwd)/$(basename "$f")" >>"$list"
+done
+
+ffmpeg -y -f concat -safe 0 -i "$list" -c copy "$out"
+echo "Wrote $out ($(du -h "$out" | cut -f1))"

demos/blog-to-podcast/demo.json

@@ -0,0 +1,6 @@
+{
+  "order": 30,
+  "title": "Blog post to podcast",
+  "stack": "TypeScript (SDK)",
+  "blurb": "Point it at any blog post URL, get a podcast-style MP3 out: fetches the article, extracts the text, chunks on sentence boundaries, synthesizes each chunk via the Speechify API, and stitches the parts together with ffmpeg."
+}

demos/blog-to-podcast/package.json

@@ -0,0 +1,20 @@
+{
+  "name": "blog-to-podcast",
+  "version": "0.1.0",
+  "private": true,
+  "type": "module",
+  "description": "Turn a blog post into a podcast episode with the Speechify API",
+  "scripts": {
+    "start": "tsx src/index.ts",
+    "typecheck": "tsc --noEmit"
+  },
+  "dependencies": {
+    "@speechify/api": "^3.0.1",
+    "dotenv": "^16.4.5"
+  },
+  "devDependencies": {
+    "@types/node": "^22.0.0",
+    "tsx": "^4.19.0",
+    "typescript": "^5.6.0"
+  }
+}

demos/blog-to-podcast/src/index.ts

@@ -0,0 +1,160 @@
+import "dotenv/config";
+import fs from "node:fs";
+import path from "node:path";
+import { fileURLToPath } from "node:url";
+import { SpeechifyClient, SpeechifyError } from "@speechify/api";
+
+const MAX_CHARS = 4_000;
+
+function assertKey(): string {
+  const token = process.env.SPEECHIFY_API_KEY;
+  if (!token) throw new Error("Set SPEECHIFY_API_KEY (copy .env.example to .env).");
+  return token;
+}
+
+function stripDeep(node: string, start: string, end: string): string {
+  let out = node;
+  let i = out.indexOf(start);
+  while (i !== -1) {
+    const j = out.indexOf(end, i + start.length);
+    if (j === -1) break;
+    out = out.slice(0, i) + out.slice(j + end.length);
+    i = out.indexOf(start, i);
+  }
+  return out;
+}
+
+export function extractArticleText(html: string): string {
+  const lower = html.toLowerCase();
+  let top = lower.indexOf("<article");
+  if (top === -1) top = lower.indexOf("<main");
+  if (top === -1) top = 0;
+  let body = top > 0 ? html.slice(top) : html;
+
+  for (const [s, e] of [
+    ["<script", "</script>"],
+    ["<style", "</style>"],
+    ["<nav", "</nav>"],
+    ["<aside", "</aside>"],
+    ["<footer", "</footer>"],
+    ["<form", "</form>"],
+    ["<svg", "</svg>"],
+  ]) {
+    body = stripDeep(body, s, e);
+  }
+
+  const text = body
+    .replace(/<br\s*\/?>\s*/gi, "\n")
+    .replace(/<\/(p|h1|h2|h3|h4|h5|h6|li|div|section)>/gi, "\n")
+    .replace(/<[^>]+>/g, " ")
+    .replace(/&nbsp;/gi, " ")
+    .replace(/&amp;/gi, "&")
+    .replace(/&quot;/gi, '"')
+    .replace(/&#39;|&apos;/gi, "'")
+    .replace(/&lt;/gi, "<")
+    .replace(/&gt;/gi, ">")
+    .replace(/\s+\n/g, "\n")
+    .replace(/[ \t]+/g, " ")
+    .replace(/\n{3,}/g, "\n\n")
+    .trim();
+
+  return text;
+}
+
+export function chunkText(text: string, maxLen: number = MAX_CHARS): string[] {
+  const chunks: string[] = [];
+  let buf = "";
+  const paragraphs = text.split("\n\n").map((p) => p.trim()).filter(Boolean);
+
+  for (const para of paragraphs) {
+    if (buf.length + para.length + 2 <= maxLen) {
+      buf = buf ? `${buf}\n\n${para}` : para;
+      continue;
+    }
+    if (buf) {
+      chunks.push(buf);
+      buf = "";
+    }
+    if (para.length > maxLen) {
+      for (const sent of para.split(/(?<=[.!?])\s+/)) {
+        if (buf.length + sent.length + 1 <= maxLen) {
+          buf = buf ? `${buf} ${sent}` : sent;
+        } else {
+          if (buf) chunks.push(buf);
+          buf = sent;
+        }
+      }
+    } else {
+      buf = para;
+    }
+  }
+  if (buf) chunks.push(buf);
+  return chunks;
+}
+
+async function synthesizeChunks(client: SpeechifyClient, chunks: string[], voiceId: string, model: string, outDir: string): Promise<number> {
+  fs.mkdirSync(outDir, { recursive: true });
+  let total = 0;
+  for (let i = 0; i < chunks.length; i++) {
+    const resp = await client.audio.speech({
+      input: chunks[i],
+      voice_id: voiceId,
+      audio_format: "mp3",
+      model: model as "simba-english" | "simba-multilingual" | "simba-3.0" | "simba-3.2",
+    });
+    const out = path.join(outDir, `part-${String(i).padStart(3, "0")}.mp3`);
+    fs.writeFileSync(out, Buffer.from(resp.audio_data, "base64"));
+    total += resp.billable_characters_count ?? chunks[i].length;
+    console.log(`  wrote ${out} (${resp.billable_characters_count} billable chars)`);
+  }
+  return total;
+}
+
+async function main() {
+  const url = process.argv[2];
+  const voiceId = process.argv.includes("--voice") ? process.argv[process.argv.indexOf("--voice") + 1] : "george";
+  const model = process.argv.includes("--model") ? process.argv[process.argv.indexOf("--model") + 1] : "simba-english";
+
+  if (!url) {
+    console.error("Usage: npm start -- <blog-post-url> [--voice <id>] [--model <id>]");
+    process.exit(1);
+  }
+  if (!/^https?:\/\//i.test(url)) {
+    console.error(`Not a URL: ${url}`);
+    process.exit(1);
+  }
+
+  const client = new SpeechifyClient({ token: assertKey() });
+
+  console.log(`Fetching ${url} ...`);
+  const res = await fetch(url, { headers: { "user-agent": "speechify-blog-to-podcast-demo" } });
+  if (!res.ok) {
+    throw new Error(`Fetch failed: ${res.status} ${res.statusText} for ${url}`);
+  }
+  const html = await res.text();
+  const text = extractArticleText(html);
+  console.log(`Extracted ${text.length.toLocaleString()} characters of article text.`);
+  if (!text) {
+    throw new Error("No readable article text found — the page may be JS-rendered; point this at the raw HTML or a static post.");
+  }
+
+  const chunks = chunkText(text);
+  console.log(`Split into ${chunks.length} chunk(s) (cap ${MAX_CHARS.toLocaleString()} chars).`);
+
+  const slug = new URL(url).hostname.replace(/^www\./, "") + "-" + Date.now();
+  const outDir = path.join("output", slug);
+  const total = await synthesizeChunks(client, chunks, voiceId, model, outDir);
+  console.log(`\nSynthesized ${chunks.length} chunks (${total.toLocaleString()} billable chars) to ${outDir}/`);
+  console.log(`Stitch into one episode: ./concat.sh ${outDir} podcast-episode.mp3`);
+}
+
+if (process.argv[1] && fileURLToPath(import.meta.url) === path.resolve(process.argv[1])) {
+  main().catch((err) => {
+    if (err instanceof SpeechifyError) {
+      console.error(err.message);
+    } else {
+      console.error(err instanceof Error ? err.message : err);
+    }
+    process.exit(1);
+  });
+}

demos/blog-to-podcast/tsconfig.json

@@ -0,0 +1,14 @@
+{
+  "compilerOptions": {
+    "target": "ES2022",
+    "module": "ESNext",
+    "moduleResolution": "Bundler",
+    "strict": true,
+    "esModuleInterop": true,
+    "skipLibCheck": true,
+    "resolveJsonModule": true,
+    "noEmit": true,
+    "types": ["node"]
+  },
+  "include": ["src/**/*.ts"]
+}