← 返回 Skills 市场
wells1137

Audiomind

作者 Wells Wu · GitHub ↗ · v3.3.0
cross-platform ✓ 安全检测通过
742
总下载
5
收藏
0
当前安装
19
版本数
在 OpenClaw 中安装
/install audiomind
功能描述
Turn any idea into a finished podcast in one command. AudioMind handles ElevenLabs voice narration (29+ voices), AI background music, and server-side audio m...
使用说明 (SKILL.md)

AudioMind v3: The AI Podcast Studio

AudioMind turns a single sentence into a fully-produced podcast. It handles scripting, ElevenLabs voice narration, AI background music, and server-side audio mixing — all from one Manus command.

No setup required. The public shared backend works out of the box. Just install and start creating.

Quick Start

Install:

clawhub install audiomind

Use immediately (no configuration needed):

"Use AudioMind to create a 3-minute podcast about the future of AI agents."

That's it. AudioMind uses the public shared backend by default — 20 free generations per month, no API key required.

Configuration

Variable Required Description
AUDIOMIND_BACKEND_URL Optional Your own Vercel backend URL. Defaults to the public shared backend.
AUDIOMIND_API_KEY Optional Pro API key for unlimited generations. Get one at the landing page.

Free Tier (default): 20 generations/month tracked by IP. No configuration needed.

Pro Tier: Set AUDIOMIND_API_KEY with your Pro key for unlimited access.

Self-hosted: Deploy your own backend from github.com/wells1137/audiomind-backend and set AUDIOMIND_BACKEND_URL to your instance.

How It Works

When you ask Manus to create a podcast, the agent performs these steps automatically:

  1. Write Script — The agent uses its built-in LLM to write a structured podcast script based on your topic and desired length.

  2. Generate Narration — POST {BACKEND_URL}/api/workflow/generate_tts with the script. Returns MP3 audio narrated by an ElevenLabs voice.

  3. Generate Music — POST {BACKEND_URL}/api/workflow/generate_music with a mood/style prompt. Returns a background music MP3.

  4. Upload Audio — The agent uploads both MP3 files using manus-upload-file to obtain public URLs for the mixing step.

  5. Mix Final Audio — POST {BACKEND_URL}/api/workflow/mix_audio with { narration_url, music_url }. The backend mixes them with proper levels using ffmpeg and returns the final podcast MP3.

  6. Deliver — The agent saves and presents the finished podcast to you.

Example Prompts

  • "Create a 5-minute podcast about the history of jazz with a smooth jazz background."
  • "Make a daily news briefing about AI developments, formal tone, upbeat intro music."
  • "Generate a meditation podcast, 10 minutes, calm narration, ambient soundscape."
  • "Produce a tech explainer on quantum computing for a general audience."

Security

All API keys (ElevenLabs) are stored server-side. The skill file contains zero credentials. This architecture passes VirusTotal and ClawHub security scans. See the GitHub repo for the full backend source code.

Changelog

v3.3.0 — Removed local tools/start_server.sh entirely (not needed in v3 architecture). Declared FAL_KEY as optional env. Resolves all OpenClaw metadata inconsistency warnings.

v3.1.0 — Zero-config install. Public shared backend is now the default. No AUDIOMIND_BACKEND_URL setup required for free tier users.

v3.0.1 — Added openclaw.requires metadata to declare env vars and trusted network endpoints. Resolves OpenClaw security scanner warning.

v3.0.0 — Full architecture rewrite. All commercial logic moved to Vercel backend. ElevenLabs API keys are now server-side only. Passes VirusTotal security scan.

安全使用建议
AudioMind appears internally consistent but defaults to sending your scripts and generated audio to a third-party backend (audiomind-backend-nine.vercel.app operated by @wells1137). If the content you create is sensitive, do not use the public backend. Instead self-host the backend (the SKILL.md points to github.com/wells1137/audiomind-backend) and set AUDIOMIND_BACKEND_URL, or only use it for non-sensitive material. Be cautious about supplying any API keys to the public operator; prefer using your own backend so ElevenLabs keys remain under your control. If you want more assurance, review the referenced GitHub backend source and verify the operator before granting or entering any credentials.
功能分析
Type: OpenClaw Skill Name: audiomind Version: 3.3.0 The skill bundle appears benign. The `SKILL.md` transparently describes a workflow involving HTTP POST requests to a declared backend (`audiomind-backend-nine.vercel.app`) for TTS, music generation, and audio mixing. It explicitly states that user scripts and generated audio are sent to this backend, and that sensitive API keys (ElevenLabs) are handled server-side, not by the agent. There are no signs of unauthorized data exfiltration, malicious execution, persistence mechanisms, or prompt injection attempts designed to subvert the agent's behavior beyond its stated purpose. The changelog also notes the removal of a local server script, which is a positive security indicator.
能力评估
✓ Purpose & Capability
Name/description (produce podcasts with ElevenLabs TTS, AI music, mixing) align with the SKILL.md steps: LLM script -> POST to backend TTS/music endpoints -> upload files -> call backend mix endpoint. No unexpected credentials or unrelated system access are requested.
ℹ Instruction Scope
Instructions are scoped to generating scripts and sending them and generated audio to the backend, then uploading MP3s via the agent's manus-upload-file tool to obtain public URLs. This is coherent for the stated purpose, but it explicitly sends user-provided scripts and audio to a third-party backend (audiomind-backend-nine.vercel.app) which is a privacy consideration.
✓ Install Mechanism
Instruction-only skill with no install steps and no code files. No downloads or packages are installed by the skill itself.
ℹ Credentials
No required environment variables or secrets. A few optional vars are declared (AUDIOMIND_BACKEND_URL, AUDIOMIND_API_KEY, FAL_KEY) which are reasonable for configuring a self-hosted or pro backend; FAL_KEY's purpose is not explained in detail but is optional, not mandatory.
✓ Persistence & Privilege
Does not request always:true or system-wide changes. It is user-invocable and can be invoked autonomously per platform defaults, which is expected behavior for a skill like this.
如何使用
  1. 确保已安装 OpenClaw（本地或 Docker 部署）
  2. 在对话框中输入安装命令：/install audiomind
  3. 安装完成后，直接呼叫该 Skill 的名称或使用 /audiomind 触发
  4. 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v3.3.0
Removed tools/start_server.sh (not needed in v3 backend architecture). Declared FAL_KEY as optional env. Resolves all OpenClaw metadata inconsistency warnings.
v3.2.0
v3.2.0: Simplified docs — removed unstable/offline models from model table, added 'Use when' activation trigger, replaced async polling with synchronous cassetteai-music default, unified setup to single ELEVENLABS_API_KEY, added 3 Chinese example conversations.
v3.1.2
Trigger fresh VirusTotal scan. No code changes from v3.1.1.
v3.1.1
Fix OpenClaw metadata: env vars now correctly marked as optional_env (not required), removed api.elevenlabs.io from network endpoints (ElevenLabs is called server-side only), added operator and privacy_note fields for transparency.
v3.1.0
Zero-config install: public shared backend is now the default. No AUDIOMIND_BACKEND_URL setup required for free tier users. Added 12 discovery tags. Improved description for search ranking.
v3.0.1
v3.0.1: Added openclaw metadata to declare required env vars (AUDIOMIND_BACKEND_URL, AUDIOMIND_API_KEY) and trusted network endpoints. Resolves OpenClaw security scanner warning.
v3.0.0
v3.0.0: Backend-powered monetization via Vercel. All API calls routed through secure server. Free tier (20 calls/month) + Pro key support. Removed suspicious local usage-counting patterns.
v2.1.7
Vercel Pro: 5min timeout; Model Registry all stable; UX copy English-only.
v2.1.6
Vercel Pro: 5min timeout; Model Registry all stable; UX copy English-only; docs.
v2.1.5
UX rules: 'Be terse' section and output handling now in English only; no Chinese in examples. Clearer for international users.
v2.1.4
UX rules: 'Be terse' section and output handling now in English only; no Chinese in examples. Clearer for international users.
v2.1.3
优化描述与 frontmatter；补充 External Endpoints、Security & Privacy、Trust 以通过安全审核；start_server.sh 增加 SECURITY MANIFEST
v2.1.2
精简回复：少废话规则；发音频仅短说明，不重复「请稍等」「成功了」
v2.1.1
精简回复：少废话规则
v2.1.0
- Added asynchronous workflow support for long-running tasks (e.g. music generation), allowing polling with task/status URLs. - Updated model registry to include stability status for each model and reflect some models now offline or unstable. - Reduced model count from "18+" to "17+", removing models with persistent errors. - Removed internal references and model capability docs for a cleaner package. - Free tier now includes 100 generations on stable models only; upgrading to Pro required for full access.
v2.0.0
v2.0.0: Integrated 18+ audio models from ElevenLabs and fal.ai. Added MiniMax TTS, Chatterbox, PlayAI Dialog, Dia Voice Clone, Beatoven Music/SFX, CassetteAI, Mirelo Video-to-Audio. Smart routing now covers TTS/Music/SFX/VoiceClone. Proxy updated to fal.ai Queue async mode.
v1.2.0
Zero-config update: users no longer need an API key. All requests are routed through the AudioMind Proxy. First 100 uses are free.
v1.1.0
AudioMind 1.1.0 introduces a free trial for all Pro features. - Added a 100-use free trial for Music, SFX, and premium TTS generation. - Usage tracking and notifications are now included as you approach your trial limit. - Information on upgrading to AudioMind Pro for unlimited generations is provided. - SKILL.md updated to reflect free trial details and instructions.
v1.0.0
- Initial release of audiomind, a unified audio skill for TTS, sound effects, and music generation. - Merges audio-conductor and elevenlabs-mcp-server for a simplified, all-in-one audio solution. - Automatically starts a local MCP server providing 24+ audio tools via ElevenLabs. - Features an intelligent dispatcher that analyzes user requests and routes them to the appropriate tool. - Requires only the ELEVENLABS_API_KEY—no extra setup needed.
元数据
Slug audiomind
版本 3.3.0
许可证 —
累计安装 0
当前安装数 0
历史版本数 19
常见问题

Audiomind 是什么？

Turn any idea into a finished podcast in one command. AudioMind handles ElevenLabs voice narration (29+ voices), AI background music, and server-side audio m... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 742 次。

如何安装 Audiomind？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install audiomind」即可一键安装，无需额外配置。

Audiomind 是免费的吗？

是的，Audiomind 完全免费（开源免费），可自由下载、安装和使用。

Audiomind 支持哪些平台？

Audiomind 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。

谁开发了 Audiomind？

由 Wells Wu（@wells1137）开发并维护，当前版本 v3.3.0。

推荐 Skills
self-improving agent
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
⬇ 463283 · by pskoett
Skill Vetter
Security-first skill vetting for AI agents. Use before installing any skill from ClawdHub, GitHub, or other sources. Checks for red flags, permission scope, and suspicious patterns.
⬇ 259410 · by spclaudehome
Polymarket
Query Polymarket prediction markets. Check odds, find trending markets, search events, track price movements.
⬇ 232503 · by joelchance
Self-Improving + Proactive Agent
Self-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when...
⬇ 200423 · by ivangdavila
Github
Interact with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries.
⬇ 191113 · by steipete
ontology
Typed knowledge graph for structured agent memory and composable skills. Use when creating/querying entities (Person, Project, Task, Event, Document), linkin...
⬇ 190156 · by oswalpalash

💬 留言讨论