Top

Wapto documentation

AI Models Management

Welcome to WAPTO!

Get an AI summary of this page
ChatGPT Gemini Grok Perplexity

Overview

The **AI Models** module is the intelligence orchestration layer of the Wapto platform. This feature enables administrators to integrate, configure, and swap Large Language Models (LLMs) from premier providers—including **OpenAI, Google Gemini, DeepSeek, and xAI**—to power platform-wide automations, smart chat assistants, and content generation workflows.

By centralizing AI governance, Wapto allows you to optimize for cost, performance, and accuracy by toggling between different model tiers (e.g., GPT-4o vs. Gemini Flash) without altering downstream business logic. Advanced configurations like custom headers and fine-tuning parameters ensure your AI interactions are secure and precisely calibrated.

AI Portfolio Architecture

The master list provides a real-time status of your integrated intelligence assets:

Metadata Field Resource Description
Provider The host platform (OpenAI, Google, DeepSeek, xAI, etc.).
Model Identifier The specific API name (e.g., `gpt-4o`, `deepseek-chat`).
Base URL The API endpoint. Supports custom proxy or self-hosted overrides.
Is Default Marked 'YES' for the primary model used for global platform features.
Status Toggle Enables or disables the model's availability for the platform.

Intelligence Configuration Protocol

Navigate to System Settings → AI Models and click Add AI Model.
Select a Provider and enter a descriptive Display Name.
Enter the precise Model Name (API ID) provided by the host.
Configure Advanced Parameters (Headers, Temp, Tokens) if required.
Toggle Set as Default if this is your primary intelligence engine.
Request Authentication

Custom headers allow for secure, authenticated requests to unique provider endpoints:

  • Auth Tokens: Pass specific bearer tokens or organization IDs.
  • Custom Prototyping: Define provider-specific HTTP headers.
  • Proxy Support: Configure headers required for secure internal gateways.
Orchestration Best Practices
  • Global Default Strategy: Always maintain one 'Default' model to prevent feature failure if a specific model is deleted.
  • Base URL Overrides: Leave the Base URL empty to use the provider's official global endpoint by default.
  • Pre-Production Testing: Use the **Test Model** action (if available) before enabling for user-facing automation.
  • Credential Safety: Keep API keys in the specific Provider module; do not paste sensitive keys into the 'Description' field.
INTELLIGENCE ALERT

Inconsistent Model Identifiers or expired Base URLs will cause platform-wide AI features to enter a 'Silent Failure' state. Monitor your provider dashboards regularly for API health.