Models

The model list and prices change, so this site does not copy them. Read them from the API, or browse each model’s page at https://blon.ai/models/<model_id>.

List models

curl https://api.blon.ai/v1/models \
  -H "Authorization: Bearer $BLON_API_KEY"

Each entry in data is a model summary:

Field	Meaning
`id`	The id you pass as `model`
`name`	Display name
`category`	Output and task, such as `image.edit` or `video.generate`
`availability`	`operational`, or `disabled` while the model is stopped
`output`	`type` (`image`, `video` or `audio`) and `content_type`
`pricing`	The current price, in your currency
`price_until`	When a limited-time price ends; `null` when it has no end
`docs_url`	A page about the model

The list shows only models offered to the caller. A model that is not offered returns 404 model_not_found, just like one that does not exist.

Read a model’s inputs

curl https://api.blon.ai/v1/models/nano-banana-2-lite \
  -H "Authorization: Bearer $BLON_API_KEY"

The single model adds inputs. Each input has an id (the key in the job’s input), required, a schema and an optional default.

`schema.type`	Value in `input`	Extra schema fields
`text`	String	`enum`, `min_length`, `max_length`
`integer`	Whole number	`enum`, `min`, `max`
`number`	Number	`enum`, `min`, `max`
`boolean`	`true` or `false`	—
`media`	`{ "asset_id": "…" }` or `{ "url": "https://…" }`	`media_type`
`list`	Array of `items` values	`items`, `min_items`, `max_items`
`voice`	A voice selection, below	`preset_model_id`, `preset_ids`, `workspace`

Voice inputs

A voice value picks either one of the model’s preset voices or a voice in your workspace:

{ "kind": "preset", "modelId": "<preset_model_id>", "presetId": "<one of preset_ids>" }

{ "kind": "workspace", "voiceId": "<voice id>" }

Schema field	Meaning
`preset_model_id`	The `modelId` to send with a preset voice. Absent when the model has no preset voices
`preset_ids`	The preset voices you can choose
`workspace`	Whether a workspace voice is allowed

The voice value keeps camelCase keys.

Pricing

{
  "currency": "USD",
  "components": [
    { "meter": "output_images", "rate": "0.04" }
  ]
}

pricing.components lists separate meters, such as output_images or output_seconds, each with a rate per unit. A component can also have:

applies_when: conditions on the inputs (input, op, value) under which the component applies.
tiers: other rates that apply when their when conditions match; rate is used otherwise.

To get the amount for a specific request, use the estimate endpoint. A finished job lists the units it used in usage, with the same meter names.

Get started

Concepts

How-to

Help

Models

List models

Read a model’s inputs

Voice inputs

Pricing

​List models

​Read a model’s inputs

​Voice inputs

​Pricing

List models

Read a model’s inputs

Voice inputs

Pricing