Skip to content

configuration.md

Latest commit

 

History

History
97 lines (64 loc) · 3.94 KB

configuration.md

File metadata and controls

97 lines (64 loc) · 3.94 KB
Raw

First-Time Configuration

Configure your engine paths, network ports, and model directories after installation. The application stores all configuration values in settings.json.

Step 1: Open the Settings Tab

Open the configuration screen in your desktop application or web browser:

  1. Open the Local LLM Server Manager desktop window, or navigate to http://localhost:5246 in your web browser.
  2. Click the Settings tab on the top navigation bar.

Step 2: Auto-Detect Installed Engines

Use the built-in discovery service to locate installed tools automatically:

  1. Locate the Engine Detection panel in the Settings tab.
  2. Click Auto-Detect Installed Tools.

The discovery service scans your system drives, environment variables, and standard folder locations:

  • Ollama: Checks %LOCALAPPDATA%\Programs\Ollama\ollama.exe and ~/.ollama/models.
  • Stable Diffusion Forge: Checks webui-user.bat, webui.sh, and models/Stable-diffusion folders.
  • ComfyUI: Checks run_nvidia_gpu.bat, main.py, and models/ folders.
  • Kokoro TTS Engine: Checks Python virtual environments and audio scripts.

Note

Auto-detection only updates empty or invalid path entries. The scanner does not overwrite existing valid configurations.

Inspect Path Status Badges

Inspect the color-coded badge next to each configured path:

  • 🟢 Valid: The executable file exists or the target folder is accessible.
  • 🔴 Missing: The specified target path does not exist on disk.
  • ⚪ Unset: The path is empty. The system uses standard system fallbacks.

Step 3: Configure Network Ports

Verify and customize the network port numbers for each supervised engine.

Engine Service Default Port Description
Local LLM Server Manager 5246 Unified reverse proxy, MCP server, and dashboard.
Ollama LLM Engine 11434 Language model inference server.
Stable Diffusion Forge 7860 Image generation web server.
ComfyUI Engine 8188 Node-based workflow and 3D generation server.
Kokoro TTS Engine 8880 Local speech synthesis API server.

To change an engine port number:

  1. Click the port input field for the target engine.
  2. Type the new port number.
  3. Click Save Settings.

Warning

Ensure no other application uses the assigned port numbers. Port conflicts prevent AI engines from starting.

Step 4: Configure Storage Directories

Set custom storage paths to organize your model weights, checkpoints, and generated media files.

  1. Locate the Storage Locations section in the Settings tab.
  2. Set the directory paths for each engine:
    • Ollama Models Directory: Folder containing Ollama model manifests and layer blobs.
    • Forge Models Directory: Root folder containing checkpoints, LoRAs, and VAE weights.
    • ComfyUI Models Directory: Root folder containing diffusion UNETs, text encoders, and 3D models.
    • Media Output Directory: Destination folder for generated images, videos, audio, and 3D files.
  3. Click Browse to select a directory with the system folder dialog, or type the absolute path manually.
  4. Click Save Path to confirm your changes.

Tip

Select a fast NVMe solid-state drive for your model directories. Fast disk access speeds up model preloading.

Step 5: Configure Modular Feature Packs

Enable or disable optional capabilities through the Component Manager in the Settings tab:

  • Video Feature Pack (ext_video): Installs ComfyUI workflow presets for Wan 2.2, LTX-2.5, and HunyuanVideo.
  • Audio Feature Pack (ext_audio): Installs Kokoro TTS scripts, Stable Audio presets, and YuE music workflows.
  1. Scroll to the Component Manager panel.
  2. Click Install next to the desired feature pack.
  3. Wait for the component installer to complete the process.

Next Steps

Proceed to the Quickstart Guide to download your first model and run your first prompt.