Skip to content
modelBridge Docs

How modelBridge handles new models

How modelBridge handles new and unknown models

Section titled “How modelBridge handles new and unknown models”

fal.ai adds new models every week. modelBridge is built so that new models work immediately — without a plugin update, without manual configuration, and without broken UI.

Automatic interface generation

Section titled “Automatic interface generation”

modelBridge never hardcodes a layout for a specific model. Instead, every model’s interface is built automatically from the model’s own specification when you install it.

The result: every parameter a model exposes becomes an appropriate input control in the panel — sliders for numeric ranges, dropdowns for options, drop zones for media, toggles for on/off settings. New models get complete, correct interfaces without any manual work from the modelBridge team.

Contextual help for every field

Section titled “Contextual help for every field”

Every input field in modelBridge has guidance — even for parameters that did not exist when you installed the plugin.

Help is delivered through a three-tier system:

Hand-written explanations for common parameters across the fal.ai catalog. Includes recommended starting values, editor-specific context, and links to relevant Academy articles.

Automatic descriptions sourced from the model’s own documentation when no hand-written explanation exists. These go through a quality check — vague or purely technical descriptions are filtered out and replaced with something more useful.

Generic fallback for any remaining fields:

“Advanced parameter — see parameter reference for details.” with a direct link to the parameter reference.

No field is ever left blank or unexplained.

Learn more — always a path forward

Section titled “Learn more — always a path forward”

Every field with documentation has a “Learn more” link in its ⓘ tooltip — either pointing to a specific Academy article or to the relevant section of the parameter reference. There is always somewhere to go.

Graceful handling of new parameter types

Section titled “Graceful handling of new parameter types”

When modelBridge encounters a parameter type it has not seen before, it follows a strict no-silent-failure policy:

  • The field always appears — it is never dropped
  • The field is always labeled — never anonymous
  • The field always explains what format is expected
  • The field always links to further documentation

The user knows what they are looking at and what to do with it — even for parameters that are entirely new.

Fixed values stay fixed

Section titled “Fixed values stay fixed”

Some model parameters have values that are required by the model and cannot be changed. modelBridge shows these as clearly labeled read-only fields. The correct value is always sent. The user cannot accidentally provide something invalid.

What this means over time

Section titled “What this means over time”

Today: install modelBridge, access 1,000+ models, every field rendered and explained.

Next month: fal.ai adds new models. They appear in Browse automatically. New parameters get interfaces and contextual help without a plugin update.

Six months from now: same experience. The catalog has grown. modelBridge has kept up — automatically.

modelBridge is not a static integration. It is a living layer that grows with fal.ai’s catalog without requiring constant manual updates.

Section titled “Related”