Join the Discord to get installation help, report bugs, or show off the crazy conversations your pawns have!

Requires .NET 9:

• The latest release should contain all the .NET files you need.

RimDialogue Local can be run with:

• A local LLM using Ollama, or
• A cloud LLM using an API key.

To run RimDialogue with a local LLM:

1. Download and install Ollama
   Ollama Download Page

2. Installation and configuration instructions
   Ollama GitHub Repository

System Requirements:
Depending on the model you choose, Ollama may require a powerful machine.
Llama 3.2 3B and Llama 3.2 1B are OK and will run on most machines.
Find more models here: Ollama Model Library

If you have a AMD GPU you may want to do something like ollama-for-amd.

1. Go to the model page on the Ollama website and copy the model name.
2. Open a command prompt by typing cmd in the Windows search bar.
3. In the command window, type ollama pull <paste model name> and press enter. The model should now start downloading.

If you prefer uncensored models, here are some options:

• Llama 2 Uncensored
• Mistral Uncensored
• Orca2 Uncensored
• Mixtral Uncensored
• Zephyr Uncensored
• Wizard Vicuna Uncensored
• WizardLM Uncensored

If running a local LLM isn’t an option, you can use an API key for a cloud-hosted LLM.

Supported Providers:

• Groq
  Get API keys: Groq API Keys

  Groq offers a free tier, but it’s heavily throttled.

• AWS & OpenAI

  Warning: These options are complex to configure and may incur high costs if set up improperly.

1. Go to the RimDialogue Server Releases.
2. Download the latest RimDialogueLocalServer_<version>.zip.
3. Unzip to a directory of your choice.

1. Open the appsettings.json file in a text editor.

2. In the MODELS area, configure your provider:

   • For Ollama:
     • Ensure OllamaUrl points to the correct port (default: 11434).
     • Set OllamaModelId to your chosen model (e.g., "llama3.2").
   • For Cloud Providers:
     • Fill in your API credentials in the MODELS section for your provider (e.g., GroqApiKey and GroqModelId for Groq).

3. Adjust optional settings:

   • RateLimit: Sets the number of requests per second allowed.
     • For Ollama you can set this higher (0.5 - 1.0) depending on your machine.
     • For cloud providers this will depend on your token budget. If your provider limits requests per minute, turn this down.
       • 0.016667 is 1 request per minute.
       • 0.166667 is 10 requests per minute.
       • 0.416667 is 25 requests per minute.
   • MaxPromptLength: Limits prompt size before truncation.
     • Set lower for tight input token budgets or higher (40,000–50,000) for local setups.
   • Options Settings:
     • Enable / disable additional prompt data with boolean options under //OPTIONS SETTINGS.
     • If your cloud provider limits input tokens, turn some of these settings off to reduce the prompt size and the number of input tokens.
   • Server Port:
     • By default, the server runs on port 7293. Change this in the Urls field if needed.

• Subscribe to the mod on Steam:
  RimDialogue Mod Page

• Alternatively, launch RimWorld, press Mods, and use the Steam Workshop browser to subscribe.

1. From RimWorld’s main menu, press Mods.
2. Find RimDialogue in the left column.
3. Click Enable in the mod description window.
4. Press Save and apply changes.

1. From the main menu, press Options.
2. Select Mod Options.
3. Choose RimDialogue.
4. Scroll to the Server URL setting and set it to:
   http://localhost:7293/

   Adjust the port if you changed it during server configuration.

5. Restart Rimworld.
6. Go back into the RimDialogue mod options and select the model you configured with the MODEL button.

1. Run RimDialogueLocal.exe from the installation folder.
2. Start your RimWorld game normally.

Contributions are welcome! Feel free to open issues or pull requests.

This project is licensed under the CC BY-NC-SA 4.0 International License.