Skip to main content
Glama

Erase region

erase_region

PROJECT-SCOPED: this call acts only on the explicit project_id and returns the project identity with its result. TRULY REMOVE whatever is inside a rectangle — repaints those pixels and reconstructs the background, so the thing is GONE, not covered. Use it for a word, a sign, a sticker, a logo, a person's name on screen, or any object the user wants taken out. SEVERAL marks (a watermark AND a handle AND a caption bar) go in ONE call as regions=[{x,y,w,h,fill?,start?,end?}, ...]. The repaint costs time proportional to the WINDOW you erase, not the video — so pass start/end around when the mark is actually visible and the erase lands in seconds; earlier erases are never redone. x,y = TOP-LEFT corner, w,h = size, all FRACTIONS (0-1) of the SOURCE frame — get them from find_burned_text rather than estimating. fill: 'text' (default — repaints only the letter strokes and keeps the picture behind them; best for captions/handles) or 'box' (repaints the whole rectangle; use for an OBJECT or a solid graphic). start/end (SOURCE seconds) limit it to a window; omit both for the whole video. Reconstruction is excellent for thin text and for anything on a steady shot; a large object on a moving, detailed background can leave a soft patch — the result is measured and reported back to you, so check it before you promise anything.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
hNo
wNo
xNo
yNo
endNo
fillNo
startNo
regionsNo
project_idYesRequired immutable scope for this call. Copy the id from list_projects/open_project/project_state; the active-project pointer is never used to guess.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A3.9/5.0
Behavior1/5

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

The description claims the tool 'TRULY REMOVE[s]' content, which is a destructive operation, but the annotations declare destructiveHint: false. This directly contradicts the annotation. Since the description contradicts the structured metadata, it scores 1 per the rules, even though the description is otherwise rich with behavioral detail (repaint, reconstruction, performance, return of project identity). The contradiction is flagged.

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

Conciseness4/5

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

The description is long but every sentence serves a purpose—scoping, behavior, use cases, parameter details, performance, and limitations. It is front-loaded with the most critical information ('PROJECT-SCOPED' and 'TRULY REMOVE'). It is well-structured with clear punctuation and logical flow. It is slightly verbose but not wasteful, so a 4 rather than 5.

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

Completeness5/5

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

For a complex tool with 9 parameters, no output schema, and low schema coverage, the description is remarkably complete. It covers all parameters, explains performance trade-offs, advises on coordinate sourcing, discloses limitations, and notes that the result is measured and reported back. An agent has everything needed to decide when to use it and how to invoke it correctly.

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

Parameters5/5

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

Schema description coverage is only 11% (only project_id has a description), so the description must carry the burden of parameter semantics. It does so excellently: x,y,w,h are defined as top-left corner and fractions of the source frame; fill is explained with 'text' vs 'box' meanings; start/end are described as source-second windows; and the regions array is explained. It even tells the agent to get coordinates from find_burned_text. This fully compensates for the sparse schema.

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?

The description explicitly states the tool 'TRULY REMOVE whatever is inside a rectangle — repaints those pixels and reconstructs the background, so the thing is GONE, not covered.' It names specific use cases (word, sign, sticker, logo, person's name) and distinguishes itself from siblings like blur_region or erase_burned_text by emphasizing true removal rather than covering. This is a clear, specific verb+resource with strong differentiation.

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?

The description provides clear usage context: 'Use it for a word, a sign, a sticker, a logo, a person's name on screen, or any object the user wants taken out.' It also advises getting coordinates from find_burned_text rather than estimating, and explains that multiple marks go in one call. It mentions limitations (large objects on moving backgrounds may leave a soft patch) but does not explicitly name alternative tools or say 'do not use for X', so it falls short of a 5.

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

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.