LLMs

Bring your own model. Or run it local.

VizPane's smart touches — the analyst narrative, chart suggestions, infographic copy — run on a model you choose. Connect any OpenAI-compatible provider, or keep it entirely on your own machine. And if no model is set, everything still works on a deterministic fallback.

same input · two models
The same dataset run through two different models
Illustrative example · the same data, answered by two different models.
Your choice

Any model

OpenAI-compatible APIs, or a local runtime — pick the model that fits your budget, privacy, and taste.

Private

Keep it on-box

Run a fully local model and nothing leaves your infrastructure. Off-box providers are opt-in and clearly flagged.

Resilient

Always a fallback

No model configured, or a call fails? VizPane falls back to a deterministic result — features never break.

Talk through your setup → For agents