aseprite-ai-artist

Overview Schema Source Related Servers Discussions Admin

Transform

transform

Apply translate, flip, rotate, scale, outline, or crop to content on a single cel (one layer and frame) in Aseprite. Non-90° rotations and non-integer scales require lossy mode.

Instructions

Move, flip, rotate or scale pixels on ONE cel — the target layer's image on the target frame. Ops: 'translate', 'flip', 'rotate', 'scale', 'outline', 'crop_to_content'. To transform part of a cel, crop the region out with draw op 'blit' first; there is no selection-scoped transform. Rotation is only clean at 90° multiples — arbitrary angles destroy pixel art, so anything else needs allowLossy. Scaling is nearest-neighbour and integer-only for the same reason.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`dx`	No
`dy`	No
`op`	Yes
`axis`	No	For 'flip'.
`side`	No	For 'outline'. 'outside' paints transparent pixels touching opaque ones, growing the cel by one pixel each way. 'inside' recolours opaque pixels that touch transparency instead, same size in and out.	outside
`angle`	No	Degrees, clockwise. For 'rotate'.
`color`	No	Outline colour, for 'outline'.
`frame`	No	1-based frame number. Omit to use the active frame.
`layer`	No	Layer name. Omit to use the active layer.
`factor`	No	For 'scale'.
`sprite`	No	Which sprite: its id as '#7' (what every result reports back, and the only unambiguous form — two unsaved documents are both called 'Sprite'), its filename, or its display name. Omit to use the sprite Aseprite has focused.
`diagonals`	No	For 'outline'. Treat diagonal neighbours as touching too (8-neighbour instead of 4).
`thickness`	No
`allowLossy`	No	Permit a non-90° rotation or non-integer scale. Ask the user first.

Output Schema

TableJSON Schema

Name	Required	Description	Default
`op`	Yes
`bounds`	No
`sprite`	Yes
`pixelsChanged`	Yes

Schema Changelog

Changes observed during successful MCP inspections.

Changed2 schema fields changedv0.4.0Sep 30, 2026

addedInput schema / properties / diagonals

Added value: +{
+  "default": false,
+  "description": "For 'outline'. Treat diagonal neighbours as touching too (8-neighbour instead of 4).",
+  "type": "boolean"
+}

addedInput schema / properties / side

Added value: +{
+  "default": "outside",
+  "description": "For 'outline'. 'outside' paints transparent pixels touching opaque ones, growing the cel by one pixel each way. 'inside' recolours opaque pixels that touch transparency instead, same size in and out.",
+  "enum": [
+    "outside",
+    "inside"
+  ],
+  "type": "string"
+}

Changed4 schema fields changedv0.1.6Sep 9, 2026

removedInput schema / properties / scope

Removed value: -{
-  "default": "selection",
-  "description": "What the transform applies to. 'selection' falls back to the cel when nothing is selected.",
-  "enum": [
-    "selection",
-    "cel",
-    "layer",
-    "sprite"
-  ],
-  "type": "string"
-}

changedInput schema / properties / sprite / description

Previous value: -"Sprite filename or id. Omit to use the sprite Aseprite has focused."New value: +"Which sprite: its id as '#7' (what every result reports back, and the only unambiguous form — two unsaved documents are both called 'Sprite'), its filename, or its display name. Omit to use the sprite Aseprite has focused."

removedOutput schema / properties / scope

Removed value: -{
-  "type": "string"
-}

changedOutput schema / required

Previous value: -[
-  "sprite",
-  "op",
-  "scope",
-  "pixelsChanged"
-]New value: +[
+  "sprite",
+  "op",
+  "pixelsChanged"
+]

First observedv0.1.0Sep 8, 2026

TDQS

A4.4/5.0

Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare the mutation/read profile, and the description adds meaningful behavioral context on top: rotation is clean only at 90° multiples, arbitrary angles 'destroy pixel art' and require `allowLossy`, and scaling is nearest-neighbour and integer-only. This is genuinely useful operational disclosure beyond the annotation trio. It does not describe output or failure modes, keeping it from a 5.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Three sentences, front-loaded with the verb/resource/scope, then the op list, then the two constraints. Every clause carries information (no selection transform, lossy gating, integer scaling) with no filler.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 14-parameter mutation tool with an output schema present, the description covers scope, ops, lossy caveats and the partial-transform workaround adequately. Minor gaps remain, e.g. how translation handles out-of-bounds pixels (clip vs wrap) and the default outline color, but nothing blocking correct invocation.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 71%, so the schema documents most parameters (axis, side, angle, color, factor, sprite, diagonals, allowLossy). The description still adds value by defining op-level semantics and the integer/90° constraints that explain why `factor` is integer-only and why `allowLossy` exists, which the schema's max/min bounds alone do not convey.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Opens with a specific verb set and resource ('Move, flip, rotate or scale pixels on ONE cel'), scopes it precisely to 'the target layer's image on the target frame', and enumerates the six ops. It also distinguishes itself from siblings by naming `draw` and ruling out any selection-scoped transform, so an agent can tell it apart from `select`/`draw` without opening schemas.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives a concrete when-not case and a workaround: partial transforms must go through `draw` op 'blit' because 'there is no selection-scoped transform'. It also gates lossy behavior behind `allowLossy` with an explicit 'Ask the user first'. No explicit statement of the ordinary when-to-use happy path, but the boundaries are clear.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server