# ReelForge Configuration Example # 复制此文件为 config.yaml 并填入你的配置 # Project name project_name: ReelForge # ==================== LLM Configuration ==================== # Simple 3-field configuration - works with any OpenAI SDK compatible LLM # # Popular choices (copy one of these): # # Qwen Max (推荐中文): # api_key: "sk-xxx" # base_url: "https://dashscope.aliyuncs.com/compatible-mode/v1" # model: "qwen-max" # # Get API key: https://dashscope.console.aliyun.com/apiKey # # OpenAI GPT-4o: # api_key: "sk-xxx" # base_url: "https://api.openai.com/v1" # model: "gpt-4o" # # Get API key: https://platform.openai.com/api-keys # # Claude Sonnet 4: # api_key: "sk-ant-xxx" # base_url: "https://api.anthropic.com/v1/" # model: "claude-sonnet-4-5" # # Get API key: https://console.anthropic.com/settings/keys # # DeepSeek (超高性价比): # api_key: "sk-xxx" # base_url: "https://api.deepseek.com" # model: "deepseek-chat" # # Get API key: https://platform.deepseek.com/api_keys # # Ollama (本地免费): # api_key: "ollama" # Any value works # base_url: "http://localhost:11434/v1" # model: "llama3.2" # # Install: https://ollama.com/download # # Then: ollama pull llama3.2 # llm: api_key: "" # Fill in your API key base_url: "" # LLM API endpoint model: "" # Model name # ==================== TTS Configuration ==================== # TTS uses Edge TTS by default (free, no configuration needed) # Optional: Override default voice in code using voice parameter # ==================== Image Generation Configuration ==================== # Image generation uses ComfyUI workflows # Workflows are auto-discovered from workflows/image_*.json files image: default: default # Default preset name (uses workflows/image_default.json) comfyui_url: http://127.0.0.1:8188 # Local ComfyUI server # runninghub_api_key: "" # Optional: RunningHub cloud API key # Prompt prefix - automatically added to all image prompts # Leave empty ("") if you don't want any prefix prompt_prefix: "Pure white background, minimalist illustration, matchstick figure style, black and white line drawing, simple clean lines" # Common examples: # prompt_prefix: "" # No prefix # prompt_prefix: "anime style, vibrant colors, cel shading" # Anime style # prompt_prefix: "watercolor painting, soft edges, artistic" # Watercolor # prompt_prefix: "photorealistic, 8k, professional photography" # Realistic # ==================== Notes ==================== # 1. LLM Configuration: # - Simple 3-field config: api_key, base_url, model # - Works with ANY OpenAI SDK compatible LLM # - Popular choices listed in comments above # - Switch LLM: just copy-paste different values from comments # - WebUI provides quick preset selection # # 2. TTS Configuration: # - No configuration needed, uses Edge TTS by default (free) # - Override voice in code: await reelforge.tts(text="...", voice="zh-CN-XiaoxiaoNeural") # # 3. Image Generation: # - Add workflow files: workflows/image_*.json # - Auto-discovered presets: workflows/image_flux.json -> preset="flux" # - Default preset: workflows/image_default.json # # 4. Ollama (Recommended for Privacy): # - FREE: No API costs # - PRIVATE: Data never leaves your machine # - Install: https://ollama.com/download # - Usage: ollama pull llama3.2 # # 5. Security: # - Never commit config.yaml to version control # - All sensitive data (API keys) should stay local