This document provides an overview of the AI providers available through Giselle’s Generator Nodes. Giselle integrates with multiple AI providers’ APIs to offer you a diverse range of capabilities for your workflows.

​
Provider Overview

​
Anthropic

Anthropic focuses on developing AI systems that are reliable, interpretable, and steerable. Their Claude models are designed with an emphasis on helpfulness, harmlessness, and honesty, making them excellent for applications requiring nuanced understanding and safe outputs.

Key models:

  • Claude 3.7 Sonnet - Latest model with advanced reasoning capabilities
  • Claude 3.5 Sonnet - Balanced model for general applications
  • Claude 3.5 Haiku - Speed-optimized model for rapid responses

​
Google AI

Google’s Gemini models represent their most capable and versatile AI systems. These models excel at multimodal tasks and offer extensive context windows, making them suitable for complex, context-rich applications and varied input types.

Key models:

  • Gemini 2.5 Pro - Comprehensive model with multimodal capabilities and web search
  • Gemini 2.0 Flash - Speed-optimized model for quick responses
  • Gemini 2.0 Flash Thinking - Reasoning-focused rapid response model

​
OpenAI

OpenAI provides a range of powerful models with particular strengths in code generation, creative content, and structured outputs. Their models feature robust function calling capabilities and strong performance across diverse domains.

Key models:

  • GPT-4o - Flagship multimodal model with comprehensive capabilities
  • GPT-4o-mini - Fast, affordable small model for focused tasks
  • o3-mini - Reasoning-optimized small model with large token limits

​
Perplexity AI

Perplexity specializes in search-augmented generation, providing models that can access and synthesize information from the web in real-time. Their models are particularly valuable for research tasks, fact-finding, and generating up-to-date content.

Key models:

  • Sonar Pro - Advanced search model with detailed responses and accurate citations
  • Sonar - Lightweight search model optimized for quick, factual answers

​
Fal AI

Fal AI focuses on image generation models that balance quality, speed, and versatility. Their models are designed for various creative and commercial applications, offering different optimization profiles to suit diverse needs.

Key models:

  • flux-pro/v1.1 - High-quality image generation with superior composition and detail
  • flux/schnell - Ultra-fast image generation for time-sensitive applications
  • stable-diffusion-v3-medium - Balanced model with enhanced typography capabilities

​
Use Cases in Giselle

​
Content Creation Workflows

  • Anthropic (Claude): Ideal for drafting nuanced business documents, creating policy content, and generating well-reasoned explanations in Giselle workflows. Claude models are particularly effective when handling sensitive subject matter or when ethical considerations are paramount.

  • Google (Gemini): Perfect for Giselle workflows involving mixed media inputs. Use Gemini models when your workflow processes documents containing text, images, and other media formats simultaneously. Their large context window makes them excellent for analyzing extensive documentation or research materials.

  • OpenAI (GPT): Excellent choice in Giselle for technical documentation workflows, code generation pipelines, and creative content production. The structured output capabilities make GPT models particularly useful when your Giselle workflow needs to generate content in specific formats or extract structured data.

  • Perplexity (Sonar): Optimal for Giselle research workflows that require up-to-date information. Integrate Sonar models when your workflow needs to incorporate recent developments, verify facts, or compile information from multiple sources with proper attribution.

​
Visual Content Workflows

  • Fal AI: Essential for Giselle workflows involving image generation tasks. Use flux-pro for high-quality marketing visuals, flux/schnell for rapid prototyping and iterative design processes, and stable-diffusion for balanced results with typography requirements.

​
Multi-Provider Workflows

One of Giselle’s strengths is the ability to combine different providers in a single workflow:

  • Use Perplexity to gather recent information, Claude to analyze ethical implications, and GPT to format the findings into structured documentation
  • Combine Gemini’s multimodal analysis with Fal AI’s image generation to create visual content based on document analysis
  • Leverage different models’ strengths in sequential nodes to optimize cost/performance at each workflow stage

​
Future Roadmap

At Giselle, we’re committed to continuously expanding our provider integrations and updating our supported models as the AI landscape evolves. We regularly assess new models and capabilities to ensure you have access to cutting-edge AI technology that best serves your needs.

We value your input on which models and providers you’d like to see added to our platform. If you have specific requests or feedback about our current offerings, please don’t hesitate to contact us.

​
Additional Resources

For detailed specifications and documentation on each provider’s models: