Skip to main content

Provider configuration

The LLM provider is selected in Settings → AI.

Ollama (local)​

Ollama runs locally on your machine.

Advantages:

  • 100% private — your data never leaves your machine
  • No recurring cost
  • Works offline

Constraints:

  • Requires a machine with a GPU (or patience on CPU)
  • "8B" models run comfortably on 8 GB of VRAM; "70B" models require 40+ GB

Configuration:

  1. Install Ollama (a dedicated guide is available: Ollama)
  2. Download a model: ollama pull llama3.1:8b (or another)
  3. In DM Loremind → Settings → AI → Ollama, enter the URL (default http://localhost:11434)
  4. Select the model from the auto-detected list

The model pull can also be triggered from the DM Loremind UI.

1min.ai (cloud)​

1min.ai is an aggregator that gives access to Anthropic Claude, OpenAI GPT-4, Google Gemini, Mistral, etc. through a single subscription.

Advantages:

  • High-end models accessible without powerful hardware
  • Switch between providers without changing your subscription
  • Low latency

Constraints:

  • Monthly cost
  • Your prompts pass through the 1min.ai servers and the underlying provider (Anthropic, OpenAI, etc.)
  • Depends on your internet connection

Configuration:

  1. Create a 1min.ai account and obtain an API key
  2. In DM Loremind → Settings → AI → 1min.ai, paste the key
  3. The model list loads automatically, grouped by provider

Mistral (cloud)​

Mistral AI's API — European models, particularly good in French and at producing faithful JSON.

  1. Create an account on console.mistral.ai and generate an API key.
  2. In DM Loremind → Settings → AI → Mistral, paste the key.
  3. Select the model (e.g. mistral-large-latest).

Gemini / Google AI Studio (cloud)​

Google's Gemini models, through the Google AI Studio API.

Advantage: a free API key (no credit card), a generous quota and a very large context — handy for large imports.

  1. Go to aistudio.google.com → Get API key.
  2. In DM Loremind → Settings → AI → Gemini, paste the key.
  3. Select the model (e.g. gemini-2.0-flash).

OpenRouter (cloud)​

OpenRouter routes your requests to dozens of models (Anthropic, OpenAI, Google, Meta…) through a single key.

Advantage: free models available. To stay free, choose a model marked "free", or the openrouter/free router which selects one automatically. For imports (large documents), prefer a free model with a large context.

  1. Create an account on openrouter.ai and generate an API key.
  2. In DM Loremind → Settings → AI → OpenRouter, paste the key.
  3. Select the model.

Which model should you choose?​

For DM Loremind, in practice:

  • Simple tasks (summarizing a scene, suggesting 3 ideas): Llama 3.1 8B, Claude Haiku 3.5, GPT-4o-mini are enough
  • Narrative tasks (writing a coherent backstory): Claude Sonnet, GPT-4o give the best results
  • Long-form generation (generating an entire arc): prefer Claude Sonnet — better context adherence