← Back to Skills Marketplace
309
Downloads
0
Stars
1
Active Installs
2
Versions
Install in OpenClaw
/install noizai-chat-with-anyone
Description
Chat with any real person or fictional character in their own voice by automatically finding their speech online, extracting a clean reference sample, and ge...
Usage Guidance
This skill is internally consistent with its stated purpose, but before installing: 1) Ensure you trust the Noiz service and read its privacy policy — the skill uploads images and audio to noiz.ai. 2) Provide NOIZ_API_KEY only if you accept that external uploads will occur; the script may read ~/.noiz_api_key if the env var is not set. 3) Do not use the skill to impersonate private individuals or for deception — the SKILL.md notes this but the tool cannot enforce intent. 4) Confirm ffmpeg, yt-dlp, and the referenced tts skill are installed and available. 5) Consider using a scoped/revocable API key and monitor its usage; revoke it if you see unexpected network activity. If you want stronger assurance, manually inspect the two included scripts (they're short and straightforward) and monitor outbound connections during a test run.
Capability Analysis
Type: OpenClaw Skill
Name: noizai-chat-with-anyone
Version: 0.1.1
The skill bundle provides a legitimate workflow for voice cloning and synthesis using the Noiz AI API, yt-dlp, and ffmpeg. It includes scripts for extracting audio segments from videos based on subtitles (extract_ref_segment.py) and designing voices from images or descriptions (voice_design.py), with clear ethical guidelines and no evidence of malicious intent or data exfiltration.
Capability Assessment
Purpose & Capability
The skill claims to clone/design voices and generate TTS. It needs ffmpeg/yt-dlp to fetch and extract audio and the NOIZ API to design voices; the included scripts implement exactly those tasks. Declared primaryEnv=NOIZ_API_KEY matches the voice_design script's behavior and the SKILL.md's dependency on a tts skill.
Instruction Scope
Runtime instructions are specific: web search -> download video with yt-dlp -> parse SRT -> extract audio with ffmpeg -> call tts or Noiz voice-design. The skill requires network and filesystem access (declared). It asks the agent to upload images/audio to the Noiz API and to use vision capabilities for images; it also includes explicit ethical guidance to avoid impersonation. There are no instructions to read unrelated system files or exfiltrate arbitrary data.
Install Mechanism
There is no install spec; this is instruction+script based. Both included Python scripts run locally and call known tools (ffmpeg) and an external API (noiz.ai). No remote download/install from untrusted URLs is present in the manifest.
Credentials
Only NOIZ_API_KEY is listed as the primary credential, which is appropriate. However, the voice_design script will also look for a ~/.noiz_api_key file if the env var is not set — that config path is not declared in the metadata. The skill uploads images and reference audio to an external service (https://noiz.ai), so users should be aware they are sending potentially sensitive personal media off-device.
Persistence & Privilege
The skill is not always-enabled, does not request elevated platform privileges, and does not modify other skills or global agent settings. It writes temporary files (tmp/... and voice_id.txt) but has no persistence beyond normal runtime artifacts.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install noizai-chat-with-anyone - After installation, invoke the skill by name or use
/noizai-chat-with-anyone - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.1
- Added support for generating a matching voice from an uploaded image when the person is not recognizable from their name alone.
- Introduced two workflows: A) Voice cloning from online video (name-based), and B) Voice design using an image and description (image-based).
- Included new automation scripts `extract_ref_segment.py` (to select/extract clean audio samples) and `voice_design.py` (to generate voice profiles from photos).
- Updated usage instructions, dependencies, and expanded ethical guidelines (only use public figures, inform users of synthetic speech, decline harmful requests).
- Added `yt-dlp` as a required dependency and improved clarity of prerequisites and setup steps.
v0.1.0
Initial release – enables realistic voice chat with any person or fictional character.
- Automatically finds clear speech samples online for any specified character or person
- Extracts a clean audio reference to clone the target voice
- Uses cloned voice to generate audio replies in character
- Responds to user prompts like “我想跟xxx聊天” and similar requests
- Guides include explicit steps: disambiguation, sample selection, audio extraction, and TTS generation with provided commands
- Integrates with youtube-downloader, ffmpeg, and tts skills for end-to-end workflow
Metadata
Frequently Asked Questions
What is Chat With Anyone?
Chat with any real person or fictional character in their own voice by automatically finding their speech online, extracting a clean reference sample, and ge... It is an AI Agent Skill for Claude Code / OpenClaw, with 309 downloads so far.
How do I install Chat With Anyone?
Run "/install noizai-chat-with-anyone" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Chat With Anyone free?
Yes, Chat With Anyone is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Chat With Anyone support?
Chat With Anyone is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Chat With Anyone?
It is built and maintained by kusuriuri (@ksuriuri); the current version is v0.1.1.
More Skills