create-yss-spec 1.0.0

package/template/.pi/skills/web-video-presentation/templates/scripts/extract-narrations.ts

@@ -0,0 +1,137 @@
+/**
+ * extract-narrations.ts — collect every chapter's narration array and emit
+ * a flat segment list that the TTS pipeline can consume.
+ *
+ * Run via:
+ *   npm run extract-narrations           # writes audio-segments.json
+ *   npm run extract-narrations -- --print # also prints to stdout
+ *
+ * Reads chapter order from src/registry/chapters.ts via a simple regex
+ * (no React/CSS evaluation needed). For each chapter it dynamically
+ * imports `src/chapters/<NN>-<id>/narrations.ts` (which is React-free)
+ * and flattens to:
+ *
+ *   [
+ *     { chapter, step, text, audio: "<chapter>/<step>.mp3" },
+ *     ...
+ *   ]
+ *
+ * Step indices in the JSON are 1-indexed, matching the audio file naming
+ * convention (`public/audio/<chapter>/<N>.mp3`).
+ *
+ * Empty narration strings are skipped (silent steps don't need a TTS file).
+ */
+import { readFile, writeFile } from "node:fs/promises";
+import { existsSync } from "node:fs";
+import { resolve, dirname, join } from "node:path";
+import { fileURLToPath, pathToFileURL } from "node:url";
+
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = dirname(__filename);
+const ROOT = resolve(__dirname, "..");
+const REGISTRY_PATH = resolve(ROOT, "src/registry/chapters.ts");
+const CHAPTERS_DIR = resolve(ROOT, "src/chapters");
+const OUT_PATH = resolve(ROOT, "audio-segments.json");
+
+interface Segment {
+  chapter: string;
+  step: number;
+  text: string;
+  audio: string;
+}
+
+/** Parse `src/registry/chapters.ts` to learn chapter id order. */
+async function readChapterOrder(): Promise<{ id: string; folder: string }[]> {
+  const src = await readFile(REGISTRY_PATH, "utf8");
+  // Match: id: "..."   AND   from "../chapters/<folder>/narrations"
+  const ids: string[] = [];
+  const folders: Record<string, string> = {};
+
+  for (const m of src.matchAll(/id:\s*["']([^"']+)["']/g)) ids.push(m[1]!);
+  for (const m of src.matchAll(
+    /from\s+["']\.\.\/chapters\/([^"'\/]+)\/narrations["']/g,
+  )) {
+    // We map by import order; pair 1:1 with `ids`. Both orders are the
+    // chapter declaration order in CHAPTERS so they line up.
+    const folder = m[1]!;
+    folders[folder] = folder;
+  }
+
+  // Map id → folder by reading each chapter's narrations.ts existence.
+  // Folders are typically `<NN>-<id>`; fall back to plain `<id>` if not.
+  const result: { id: string; folder: string }[] = [];
+  for (const id of ids) {
+    const candidates = Object.keys(folders).filter((f) => f.endsWith(`-${id}`));
+    const folder = candidates[0] ?? Object.keys(folders).find((f) => f === id);
+    if (!folder) {
+      throw new Error(
+        `chapter id "${id}" registered but no matching folder found ` +
+          `under src/chapters/. Expected something like NN-${id}/narrations.ts`,
+      );
+    }
+    result.push({ id, folder });
+  }
+  return result;
+}
+
+async function loadNarrations(folder: string): Promise<unknown[]> {
+  const file = join(CHAPTERS_DIR, folder, "narrations.ts");
+  if (!existsSync(file)) {
+    throw new Error(`missing narrations.ts: ${file}`);
+  }
+  const url = pathToFileURL(file).href;
+  const mod = await import(url);
+  if (!Array.isArray(mod.narrations)) {
+    throw new Error(
+      `narrations.ts in ${folder} must export an array named "narrations"`,
+    );
+  }
+  return mod.narrations as unknown[];
+}
+
+async function main() {
+  const print = process.argv.includes("--print");
+  const order = await readChapterOrder();
+
+  const segments: Segment[] = [];
+  let silentSteps = 0;
+  for (const { id, folder } of order) {
+    const arr = await loadNarrations(folder);
+    arr.forEach((entry, i) => {
+      const step = i + 1;
+      if (typeof entry !== "string") {
+        throw new Error(
+          `chapter "${id}" step ${step}: narration must be a string ` +
+            `(got ${typeof entry}). The {text, minHoldMs} form was removed; ` +
+            `if your animation is longer than the narration, write longer ` +
+            `narration, split the step, or speed the animation up.`,
+        );
+      }
+      if (entry.trim() === "") {
+        // Silent step — no TTS needed; runtime falls back to estimate.
+        silentSteps++;
+        return;
+      }
+      segments.push({
+        chapter: id,
+        step,
+        text: entry,
+        audio: `${id}/${step}.mp3`,
+      });
+    });
+  }
+
+  await writeFile(OUT_PATH, JSON.stringify(segments, null, 2) + "\n", "utf8");
+
+  console.error(
+    `✓ extracted ${segments.length} segments from ${order.length} chapters` +
+      (silentSteps > 0 ? ` (skipped ${silentSteps} silent steps)` : ""),
+  );
+  console.error(`  → ${OUT_PATH}`);
+  if (print) console.log(JSON.stringify(segments, null, 2));
+}
+
+main().catch((err) => {
+  console.error(`✗ ${err.message ?? err}`);
+  process.exit(1);
+});

package/template/.pi/skills/web-video-presentation/templates/scripts/synthesize-audio.sh

@@ -0,0 +1,145 @@
+#!/usr/bin/env bash
+# ────────────────────────────────────────────────────────────────────
+# synthesize-audio.sh — provider-agnostic TTS runner.
+#
+# Reads audio-segments.json (produced by extract-narrations.ts) and
+# writes one mp3 per segment under public/audio/<chapter>/<N>.mp3.
+#
+# This file itself does NOT know how to call any TTS engine. It loads
+# a provider adapter from tts-providers/<name>.sh which must expose:
+#
+#   tts_synthesize <text> <out_path> [<voice>]   (required)
+#   tts_check                                    (optional)
+#   tts_install_help                             (optional)
+#
+# See tts-providers/README.md for the full contract and copy-pasteable
+# recipes for adding more providers (OpenAI / ElevenLabs / edge-tts /
+# Azure / etc.).
+#
+# Choosing a provider:
+#   PRESENTATION_TTS=<name>       env var  (default: minimax)
+#   --provider=<name>             CLI flag (overrides env)
+#
+# Choosing a voice (provider decides what's valid):
+#   PRESENTATION_TTS_VOICE=<id>   env var
+#   --voice=<id>                  CLI flag (overrides env)
+#
+# Other flags:
+#   --force                       re-synthesize even if mp3 exists
+#
+# Behavior:
+#   • Serial calls (TTS APIs commonly rate-limit parallel requests).
+#   • Skips segments whose mp3 already exists — rerun safely after a
+#     partial failure. Pass --force to re-synthesize all.
+#   • Prints progress per segment with elapsed time.
+#
+# Examples:
+#   npm run synthesize-audio
+#   npm run synthesize-audio -- --force
+#   PRESENTATION_TTS=openai npm run synthesize-audio
+#   npm run synthesize-audio -- --provider=elevenlabs --voice=Rachel
+# ────────────────────────────────────────────────────────────────────
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
+SEGMENTS="$ROOT/audio-segments.json"
+OUT_DIR="$ROOT/public/audio"
+PROVIDERS_DIR="$SCRIPT_DIR/tts-providers"
+
+PROVIDER="${PRESENTATION_TTS:-minimax}"
+VOICE="${PRESENTATION_TTS_VOICE:-}"
+FORCE=false
+
+for arg in "$@"; do
+  case "$arg" in
+    --force)         FORCE=true ;;
+    --voice=*)       VOICE="${arg#--voice=}" ;;
+    --provider=*)    PROVIDER="${arg#--provider=}" ;;
+    -h|--help)
+      sed -n '2,46p' "${BASH_SOURCE[0]}" | sed 's/^# \{0,1\}//'
+      exit 0
+      ;;
+    *) echo "✗ unknown arg: $arg" >&2; exit 1 ;;
+  esac
+done
+
+PROVIDER_FILE="$PROVIDERS_DIR/$PROVIDER.sh"
+
+# ── Pre-flight ────────────────────────────────────────────────────────
+if [[ ! -f "$SEGMENTS" ]]; then
+  echo "✗ $SEGMENTS not found. Run: npm run extract-narrations" >&2
+  exit 1
+fi
+if ! command -v jq >/dev/null; then
+  echo "✗ jq is required to read audio-segments.json" >&2
+  echo "  Install: brew install jq   (or apt-get install jq, etc.)" >&2
+  exit 1
+fi
+if [[ ! -f "$PROVIDER_FILE" ]]; then
+  echo "✗ TTS provider '$PROVIDER' not found at $PROVIDER_FILE" >&2
+  echo >&2
+  echo "  Available providers:" >&2
+  for f in "$PROVIDERS_DIR"/*.sh; do
+    [[ -f "$f" ]] || continue
+    echo "    • $(basename "$f" .sh)" >&2
+  done
+  echo >&2
+  echo "  To add your own, see $PROVIDERS_DIR/README.md" >&2
+  exit 1
+fi
+
+# shellcheck source=/dev/null
+source "$PROVIDER_FILE"
+
+if ! declare -F tts_synthesize >/dev/null; then
+  echo "✗ provider '$PROVIDER' does not define tts_synthesize" >&2
+  echo "  See $PROVIDERS_DIR/README.md for the contract." >&2
+  exit 1
+fi
+
+if declare -F tts_check >/dev/null; then
+  if ! tts_check; then
+    echo >&2
+    if declare -F tts_install_help >/dev/null; then
+      tts_install_help
+    fi
+    exit 1
+  fi
+fi
+
+# ── Main loop ─────────────────────────────────────────────────────────
+total=$(jq 'length' "$SEGMENTS")
+i=0
+synthesized=0
+skipped=0
+failed=0
+
+while IFS= read -r row; do
+  i=$((i + 1))
+  chapter=$(echo "$row" | jq -r '.chapter')
+  step=$(echo "$row" | jq -r '.step')
+  text=$(echo "$row" | jq -r '.text')
+  out="$OUT_DIR/$chapter/$step.mp3"
+
+  if [[ -f "$out" && "$FORCE" != true ]]; then
+    skipped=$((skipped + 1))
+    printf "[%3d/%d] %-20s skip (exists)\n" "$i" "$total" "$chapter/$step.mp3"
+    continue
+  fi
+
+  mkdir -p "$(dirname "$out")"
+  start=$(date +%s)
+  if tts_synthesize "$text" "$out" "$VOICE"; then
+    elapsed=$(( $(date +%s) - start ))
+    synthesized=$((synthesized + 1))
+    printf "[%3d/%d] %-20s ✓ %ss\n" "$i" "$total" "$chapter/$step.mp3" "$elapsed"
+  else
+    failed=$((failed + 1))
+    printf "[%3d/%d] %-20s ✗ FAILED\n" "$i" "$total" "$chapter/$step.mp3" >&2
+  fi
+done < <(jq -c '.[]' "$SEGMENTS")
+
+echo
+echo "✓ done (provider=$PROVIDER) — synthesized $synthesized, skipped $skipped, failed $failed"
+[[ $failed -eq 0 ]] || exit 2

package/template/.pi/skills/web-video-presentation/templates/scripts/tts-providers/README.md

@@ -0,0 +1,324 @@
+# TTS Providers
+
+`synthesize-audio.sh` 是 provider-agnostic 的 runner —— 它自己不知道
+怎么调任何 TTS，只知道循环 `audio-segments.json`、跳过已存在文件、
+打印进度。
+
+**每个 provider 是这个目录下的一个 `.sh` 文件**，定义一个
+`tts_synthesize` 函数（必需），以及可选的 `tts_check` 和
+`tts_install_help`。runner 根据 `PRESENTATION_TTS` 环境变量加载对应文件。
+
+---
+
+## 怎么用
+
+```bash
+# 默认（minimax）
+npm run synthesize-audio
+
+# 换 provider
+PRESENTATION_TTS=openai npm run synthesize-audio
+npm run synthesize-audio -- --provider=elevenlabs
+
+# 指定音色（每个 provider 自己解析）
+PRESENTATION_TTS_VOICE=alloy npm run synthesize-audio
+npm run synthesize-audio -- --voice=zh-CN-YunxiNeural
+
+# 强制全部重合成
+npm run synthesize-audio -- --force
+```
+
+`--provider` 和 `--voice` 的命令行参数会覆盖 env var。
+
+---
+
+## 内置 provider
+
+| 文件 | 后端 | 鉴权 | 备注 |
+|---|---|---|---|
+| `minimax.sh` | MiniMax `mmx` CLI | `mmx auth login --api-key` | **默认**；中文口播质量稳 |
+| `openai.sh` | OpenAI Audio Speech API | `OPENAI_API_KEY` env var | curl-based；多数 agent 已有 key |
+
+只内置这两个 —— 我们不替你做更多技术选型。其它后端的代码片段在下面，
+复制到 `tts-providers/<name>.sh` 即可启用。
+
+---
+
+## 怎么加你自己的 TTS
+
+1. 在这个目录建 `<name>.sh`（小写、kebab-case）
+2. 实现 `tts_synthesize text out_path [voice]`（必需）
+3. 可选实现 `tts_check`（启动前校验环境）和 `tts_install_help`（失败时打印怎么修）
+4. `PRESENTATION_TTS=<name> npm run synthesize-audio`
+
+---
+
+## 三函数契约
+
+### `tts_synthesize <text> <out_path> [<voice>]` （required）
+
+把一段文字写成 mp3 / 任意 web 可播的音频文件到 `<out_path>`。
+
+| 参数 | 说明 |
+|---|---|
+| `$1` | 要合成的文本（已是 UTF-8 字符串，可能包含中英文混排和标点） |
+| `$2` | 目标文件绝对路径（runner 已 `mkdir -p` 它的父目录），扩展名 `.mp3` |
+| `$3` | 音色 id（可能为空字符串，provider 自行决定默认） |
+
+成功 → exit 0 并把音频写到 `$2`。失败 → 非零退出（runner 会标 FAILED 继续下一段，不会终止全局合成）。
+
+> 如果 backend 只能出 wav / ogg，自己在函数末尾用 `ffmpeg` 转一下：
+> `ffmpeg -y -i tmp.wav -codec:a libmp3lame -qscale:a 2 "$out" >/dev/null 2>&1`
+
+### `tts_check` （optional）
+
+启动时被 runner 调一次（不是每段）。检查 CLI 是否装、API key 是否设、auth 是否通。
+未就绪 return 非零，runner 会立刻终止并打印 `tts_install_help`。
+
+### `tts_install_help` （optional）
+
+`tts_check` 失败时被 runner 调，往 stderr 打印怎么装 / 怎么登录 / 在哪拿 key。
+
+---
+
+## 常见 TTS 后端的现成片段
+
+下面**不是**内置 provider —— 是你自己写 `tts-providers/<name>.sh` 时
+可以**直接抄过去**的代码片段。复制 → 保存为 `<name>.sh` → 调通了
+就 `PRESENTATION_TTS=<name>` 用。
+
+> 大多数云 TTS 的 API key 通过环境变量传入（例如 `OPENAI_API_KEY`、
+> `ELEVENLABS_API_KEY`）。把 `export` 加到你的 shell rc，或在
+> 同目录放一个 git-ignored 的 `.env` 文件并 `set -a; source .env; set +a`。
+
+### OpenAI TTS
+
+**已内置** —— 直接看 [`openai.sh`](./openai.sh)。
+该文件也是写 HTTP-based provider 的**官方参考实现**：jq 构造 JSON
+payload、curl `-fsS` 提交、可选 base URL（接 Azure-OpenAI / 代理）、
+可选 model env var、空音色 fallback 到默认值。新接 REST API 的
+provider 直接抄它起手最快。
+
+启用：
+
+```bash
+export OPENAI_API_KEY=sk-...
+PRESENTATION_TTS=openai npm run synthesize-audio
+# 用 HD 模型 + 别的音色
+OPENAI_TTS_MODEL=tts-1-hd npm run synthesize-audio -- --provider=openai --voice=nova
+```
+
+### ElevenLabs — `tts-providers/elevenlabs.sh`
+
+```bash
+# Docs:   https://elevenlabs.io/docs/api-reference/text-to-speech
+# Env:    ELEVENLABS_API_KEY=...
+# Voice:  pass voice ID; "Rachel" default is 21m00Tcm4TlvDq8ikWAM
+# Model:  eleven_multilingual_v2 supports Chinese; eleven_turbo_v2_5 cheaper
+
+tts_check() {
+  command -v curl >/dev/null || { echo "✗ curl not found" >&2; return 1; }
+  command -v jq   >/dev/null || { echo "✗ jq not found"   >&2; return 1; }
+  [[ -n "${ELEVENLABS_API_KEY:-}" ]] || { echo "✗ ELEVENLABS_API_KEY not set" >&2; return 1; }
+}
+
+tts_install_help() {
+  cat <<'EOF' >&2
+Set your ElevenLabs key first:
+  export ELEVENLABS_API_KEY=...       # get one at https://elevenlabs.io
+EOF
+}
+
+tts_synthesize() {
+  local text="$1" out="$2" voice="${3:-21m00Tcm4TlvDq8ikWAM}"
+  local payload
+  payload=$(jq -n --arg t "$text" \
+    '{text:$t, model_id:"eleven_multilingual_v2"}')
+
+  curl -fsS -o "$out" -X POST \
+    "https://api.elevenlabs.io/v1/text-to-speech/$voice" \
+    -H "xi-api-key: $ELEVENLABS_API_KEY" \
+    -H "Content-Type: application/json" \
+    -d "$payload"
+}
+```
+
+### edge-tts — `tts-providers/edge-tts.sh`（免费 / 无 API key）
+
+```bash
+# Docs:   https://github.com/rany2/edge-tts
+# Install: pip install edge-tts
+# Voices: edge-tts --list-voices
+#   zh-CN-YunxiNeural     (男声)
+#   zh-CN-XiaoxiaoNeural  (女声)
+#   en-US-AriaNeural      (英文女声)
+#   en-US-GuyNeural       (英文男声)
+
+tts_check() {
+  command -v edge-tts >/dev/null || { echo "✗ edge-tts not found" >&2; return 1; }
+}
+
+tts_install_help() {
+  cat <<'EOF' >&2
+Install edge-tts (free, uses Microsoft Edge's TTS backend, no API key):
+  pip install edge-tts
+List available voices:
+  edge-tts --list-voices | less
+EOF
+}
+
+tts_synthesize() {
+  local text="$1" out="$2" voice="${3:-zh-CN-YunxiNeural}"
+  edge-tts --text "$text" --voice "$voice" --write-media "$out" >/dev/null 2>&1
+}
+```
+
+### macOS `say` — `tts-providers/say.sh`（离线 / 兜底）
+
+```bash
+# 系统自带，零依赖，适合 CI 跑通流程 / 离线预览。
+# 中文音色：Tingting / Sinji / Meijia（看 `say -v ?` 全列表）
+# 输出是 aiff，要 ffmpeg 转 mp3（Auto 模式 audio 标签默认认 mp3）。
+
+tts_check() {
+  command -v say     >/dev/null || { echo "✗ 'say' not available (macOS only)" >&2; return 1; }
+  command -v ffmpeg  >/dev/null || { echo "✗ ffmpeg not found (brew install ffmpeg)" >&2; return 1; }
+}
+
+tts_install_help() {
+  cat <<'EOF' >&2
+macOS-only provider. Needs ffmpeg for aiff→mp3:
+  brew install ffmpeg
+List voices:  say -v ?
+EOF
+}
+
+tts_synthesize() {
+  local text="$1" out="$2" voice="${3:-Tingting}"
+  local tmp
+  tmp=$(mktemp -t tts).aiff
+  say -v "$voice" -o "$tmp" "$text" \
+    && ffmpeg -y -i "$tmp" -codec:a libmp3lame -qscale:a 2 "$out" >/dev/null 2>&1
+  local code=$?
+  rm -f "$tmp"
+  return $code
+}
+```
+
+### Azure Speech — `tts-providers/azure.sh`
+
+```bash
+# Docs:    https://learn.microsoft.com/azure/ai-services/speech-service/rest-text-to-speech
+# Env:     AZURE_SPEECH_KEY=...   AZURE_SPEECH_REGION=eastus
+# SSML payload — Azure requires SSML, not plain JSON
+
+tts_check() {
+  command -v curl >/dev/null || { echo "✗ curl not found" >&2; return 1; }
+  [[ -n "${AZURE_SPEECH_KEY:-}"    ]] || { echo "✗ AZURE_SPEECH_KEY not set"    >&2; return 1; }
+  [[ -n "${AZURE_SPEECH_REGION:-}" ]] || { echo "✗ AZURE_SPEECH_REGION not set" >&2; return 1; }
+}
+
+tts_install_help() {
+  cat <<'EOF' >&2
+Set Azure Speech credentials:
+  export AZURE_SPEECH_KEY=...
+  export AZURE_SPEECH_REGION=eastus   # or your resource's region
+EOF
+}
+
+tts_synthesize() {
+  local text="$1" out="$2" voice="${3:-zh-CN-YunxiNeural}"
+  local lang="${voice%%-*}-${voice#*-}"; lang="${lang%%-*}-${lang#*-}"  # "zh-CN"
+  local ssml="<speak version='1.0' xml:lang='$lang'><voice xml:lang='$lang' name='$voice'>$(printf '%s' "$text" | sed 's/&/\&amp;/g; s/</\&lt;/g; s/>/\&gt;/g')</voice></speak>"
+
+  curl -fsS -o "$out" -X POST \
+    "https://${AZURE_SPEECH_REGION}.tts.speech.microsoft.com/cognitiveservices/v1" \
+    -H "Ocp-Apim-Subscription-Key: $AZURE_SPEECH_KEY" \
+    -H "Content-Type: application/ssml+xml" \
+    -H "X-Microsoft-OutputFormat: audio-24khz-48kbitrate-mono-mp3" \
+    -H "User-Agent: web-video-presentation" \
+    --data-binary "$ssml"
+}
+```
+
+### Google Cloud TTS — `tts-providers/gcloud.sh`
+
+```bash
+# Docs:   https://cloud.google.com/text-to-speech/docs/reference/rest
+# Auth:   easiest is `gcloud auth application-default login`
+#         (or set GOOGLE_APPLICATION_CREDENTIALS to a service-account json)
+# Voices: zh-CN-Wavenet-A / zh-CN-Neural2-A / en-US-Neural2-J etc.
+
+tts_check() {
+  command -v curl   >/dev/null || { echo "✗ curl not found" >&2; return 1; }
+  command -v jq     >/dev/null || { echo "✗ jq not found" >&2; return 1; }
+  command -v base64 >/dev/null || { echo "✗ base64 not found" >&2; return 1; }
+  command -v gcloud >/dev/null || { echo "✗ gcloud not found" >&2; return 1; }
+  gcloud auth application-default print-access-token >/dev/null 2>&1 || {
+    echo "✗ gcloud is not authenticated (run: gcloud auth application-default login)" >&2
+    return 1
+  }
+}
+
+tts_install_help() {
+  cat <<'EOF' >&2
+Install gcloud SDK and authenticate:
+  https://cloud.google.com/sdk/docs/install
+  gcloud auth application-default login
+  gcloud services enable texttospeech.googleapis.com
+EOF
+}
+
+tts_synthesize() {
+  local text="$1" out="$2" voice="${3:-zh-CN-Wavenet-A}"
+  local lang="${voice%-*}"; lang="${lang%-*}"  # "zh-CN"
+  local token
+  token=$(gcloud auth application-default print-access-token)
+
+  local payload
+  payload=$(jq -n --arg t "$text" --arg v "$voice" --arg l "$lang" \
+    '{input:{text:$t}, voice:{languageCode:$l, name:$v}, audioConfig:{audioEncoding:"MP3"}}')
+
+  curl -fsS -X POST https://texttospeech.googleapis.com/v1/text:synthesize \
+    -H "Authorization: Bearer $token" \
+    -H "Content-Type: application/json" \
+    -d "$payload" \
+    | jq -r '.audioContent' | base64 -d > "$out"
+}
+```
+
+---
+
+## 设计要点（自己写 provider 时记住）
+
+1. **`set -e` 友好**：runner 用 `set -euo pipefail`，所以你的函数里要么明确处理失败，要么让命令自然非零退出。不要吞错误。
+
+2. **静默成功，喧闹失败**：成功时不打印任何东西到 stdout（runner 自己打进度条）；失败时往 stderr 打详细原因。把 CLI 工具的 stdout 重定向到 `/dev/null`，stderr 留着看。
+
+3. **mp3 输出**：浏览器里 `<audio>` 标签最稳吃 mp3。能直接出 mp3 就出 mp3；非 mp3 后端在函数末尾加一步 ffmpeg。
+
+4. **音色 fallback**：`$3` 可能是空字符串。给一个合理的默认值（你最常用的中文音色 / 英文音色），不要因为没传音色就报错。
+
+5. **不要做并发**：runner 是串行的（避免 rate limit）。provider 函数也别在内部 fork 多线程。
+
+6. **不要修改全局状态**：provider 文件被 `source` 进 runner 的 shell。别 `cd`、别改 `IFS`、别 `set -e/+e` 切换，否则会污染 runner。把局部变量都 `local`。
+
+   ⚠️ 一个坑：runner 用 `set -u`，**macOS 默认 bash 3.2 在 `"${arr[@]}"` 展开空数组时会炸 `unbound variable`**。如果你的 provider 需要"可选 --voice 参数"，**不要**用 `local args=(); [[ -n $voice ]] && args=(--voice $v); cmd "${args[@]}"` —— 直接写两个 if 分支调命令（看 `minimax.sh` 的写法）。
+
+7. **API 长度上限**：单段大多数 API 都有上限（OpenAI ~4096 chars / MiniMax ~5000 / ElevenLabs ~5000）。Skill 的 narrations 单段一般 < 200 字符，正常不会撞到。如果你的 narration 撞到了，**先回去拆 step**——一个 step 的口播本来就不该这么长。
+
+---
+
+## 调试
+
+```bash
+# 看 runner 怎么调你的 provider
+bash -x scripts/synthesize-audio.sh
+
+# 跑单段试试，不动 audio-segments.json
+source scripts/tts-providers/<name>.sh
+tts_check && tts_synthesize "测试一下" /tmp/test.mp3 ""
+afplay /tmp/test.mp3   # macOS 播一下听听
+```
+
+跑通了再 `npm run synthesize-audio`。

package/template/.pi/skills/web-video-presentation/templates/scripts/tts-providers/minimax.sh

@@ -0,0 +1,50 @@
+# ────────────────────────────────────────────────────────────────────
+# MiniMax provider — uses the official mmx-cli.
+#
+# Docs:  https://platform.minimaxi.com/docs/token-plan/minimax-cli
+# Repo:  https://github.com/MiniMax-AI/cli
+#
+# Strengths: Chinese narration quality is consistently good; lots of
+# voice options; one-line CLI call.
+# ────────────────────────────────────────────────────────────────────
+
+tts_check() {
+  if ! command -v mmx >/dev/null; then
+    echo "✗ mmx CLI not found in PATH." >&2
+    return 1
+  fi
+  if ! mmx auth status >/dev/null 2>&1; then
+    echo "✗ mmx is not authenticated." >&2
+    return 1
+  fi
+}
+
+tts_install_help() {
+  cat <<'EOF' >&2
+To use the MiniMax provider:
+
+  Install:  npm install -g mmx-cli
+  Login:    mmx auth login --api-key sk-xxxxx
+            (get a key at https://platform.minimaxi.com)
+
+Or pick another provider:  PRESENTATION_TTS=<name> npm run synthesize-audio
+See tts-providers/README.md for the list and how to add your own.
+EOF
+}
+
+tts_synthesize() {
+  local text="$1"
+  local out="$2"
+  local voice="${3:-}"
+
+  # Branch instead of using an empty array — runner uses `set -u`, and
+  # macOS-default bash 3.2 fires "unbound variable" on "${arr[@]}" when
+  # arr is empty. The two-branch form is portable to old bash.
+  if [[ -n "$voice" ]]; then
+    mmx speech synthesize --voice "$voice" --text "$text" --out "$out" \
+      >/dev/null 2>&1
+  else
+    mmx speech synthesize --text "$text" --out "$out" \
+      >/dev/null 2>&1
+  fi
+}