Differences

This shows you the differences between two versions of the page.

--- mods:ai [2025/08/06 12:40] – created user
+++ mods:ai [2025/08/06 13:07] (current) – user
@@ Line 1: / Line 1: @@
-====== AI Mod ======
+====== AI Mod Setup Guide ======
-===== Open Router Provider (recommended) =====
+This page explains how to configure different AI providers for the AI Mod. Choose the provider that best fits your needs - some are free, others require payment or self-hosting.
-Some models are free to use some are paid.
-Create the key here https://openrouter.ai/settings/keys
+To select provider open Settings - AI.
-Copy the key and paste to the "API key" field in the provider settings.
-Select any model you like here https://openrouter.ai/models and copy the model id (button right after the model name). Paste the model id to the "Model" field in the porvider settings.
+===== OpenRouter (Recommended) =====
+OpenRouter allows you to access a wide selection of models via a single API.
+**Steps to set up:**
+  - Create an API key here: https://openrouter.ai/settings/keys
+  - Copy the key and paste it into the **"API key"** field in the provider settings.
+  - Browse available models here: https://openrouter.ai/models
+  - Copy the **Model ID** (click the button next to the model name) and paste it into the **"Model"** field in the provider settings.
+**Notes:**
+  * Some models are free, others are paid (via per-request credits).
+  * Many popular models are available, including GPT-4, Claude, Mistral, MythoMax, and more.
 ===== OpenAI =====
-This provider also can be used with other models that provides OpenAI-like API.
+This provider supports all models exposed through the official OpenAI API, including GPT-3.5 and GPT-4.
+**Steps to set up:**
+  - Create an API key here: https://platform.openai.com/api-keys
+  - Paste the key into the **"API key"** field in the provider settings.
+  - Use the model name (e.g. `gpt-3.5-turbo` or `gpt-4`) in the **"Model"** field.
+**Notes:**
+  * Usage is billed based on tokens (input + output).
+  * Official, stable access to OpenAI models.
+  * No NSFW content allowed — requests may be blocked or filtered.
+===== Ollama (Self-Hosted) =====
+Ollama allows you to run models locally on your own machine.
+**Steps to set up:**
+  - Download and install Ollama: https://ollama.com/
+  - Launch it and pull a model, e.g. `ollama run llama3`
+  - In the provider settings, use:
+    - **API URL:** `http://localhost:11434`
+    - **Model:** `llama3` (or any model you've pulled)
+**Notes:**
+  * Runs locally, no data is sent to third parties.
+  * You need a capable GPU for most models.
+  * Useful for full privacy and offline use.
+===== Cloudflare AI Gateway =====
+A proxy layer to route requests to other AI providers through Cloudflare infrastructure.
+**Steps to set up:**
+  - Visit the documentation: https://developers.cloudflare.com/ai-gateway/
+  - Register and configure your gateway.
+  - Use the provided endpoint and key in your provider settings.
-Create the key here https://platform.openai.com/api-keys
+**Notes:**
-Copy the key and paste to the "API key" field in the provider settings.
+  * Free tier available with some limits.
-===== Ollama =====
+===== Compatibility with OpenAI-like APIs =====
-Self-hosted AI.
+Many other providers (e.g., Together.ai, DeepInfra, Groq) support OpenAI-compatible APIs. If a provider gives you:
+  - a base URL,
+  - an API key, and
+  - a model name,
-You can get it here https://ollama.com/
+...you can usually plug those into the AI Mod using the "OpenAI" provider option.
-===== CloudFlare =====
-Free with limits.
-Documentation here https://developers.cloudflare.com/ai-gateway/
+**Important:**
+  - Some models may block or filter adult content.
+  - If using NSFW content, prefer self-hosted or OpenRouter models that explicitly allow it.