Skaggsxyz Moltstream Moltstream

MoltStream Skill

What you can do

• Deploy an AI-powered live stream on Kick with one command
• React to viewer chat in real-time via LLM (Gemini / Claude)
• Generate voiced responses (Fish Audio / ElevenLabs / OpenAI)
• Animate an avatar with lip sync and chat overlay
• Compose scenes declaratively (video, overlays, audio)
• Enforce content policies and safety guardrails
• Log every decision as a structured reasoning trace
• Generate AI characters from photos (Gemini Vision + Imagen)

Quick start

npx moltstream init     # configure your agent
npx moltstream start    # go live on Kick
npx moltstream mcp      # start MCP server for Claude/Cursor

MCP Integration

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "moltstream": { "command": "npx", "args": ["moltstream", "mcp"] }
  }
}

Available tools: get_status, start_stream, stop_stream, send_chat, get_chat_log, get_traces, update_personality, obs_control, configure

Core API

import { MoltAgent } from '@moltstream/core';

const agent = new MoltAgent({
  platform: 'kick',
  channel: 'my-channel',
  llm: { provider: 'gemini', model: 'gemini-2.5-flash' },
  tts: { provider: 'fish' },
});

agent.onChat(async (message, ctx) => {
  const response = await ctx.llm.generate(message.text);
  await ctx.tts.speak(response);
});

agent.start();

Error handling

All errors are typed and include:

• code: Machine-readable error identifier
• retryable: Boolean flag
• retryAfter: Milliseconds until retry (when applicable)
• suggestion: Human/agent-readable recovery guidance

Packages

Links

• GitHub: https://github.com/skaggsxyz/moltstream
• Website: https://moltstream.app
• npm: https://www.npmjs.com/package/moltstream
• Examples: https://github.com/skaggsxyz/moltstream/tree/main/examples

Source: skaggsxyz/moltstream — distributed by TomeVault.