Orchestrate OpenAI, Anthropic, Gemini, local models and more in a single flow, the right model for each task.
Don't marry a single model. Use GPT for reasoning, Claude for long-form analysis, a fast model for classification, each node picks its own provider, and you can chain them into a pipeline where every step plays to a model's strengths.
OpenAI, Anthropic, Gemini, Vertex, Bedrock, Groq, Mistral, managed keys or bring your own.
Run Ollama or LM Studio models for maximum privacy and control, no data leaves your infra.
One model extracts, another analyzes, a third synthesizes, chain them for more robust results.
Configure backup models that kick in if the primary is down or rate-limited, flows keep running.
Each LLM node has its own provider and model dropdown, mix freely on one canvas.
Route the output of one model into the next to build a specialized, multi-model pipeline.
Run the same input through several models to compare answers and pick the best.