Skip to main content
Glama
Hermoso

Hermoso

Official

Make video thumbnail

make_thumbnail

Create click-driving YouTube, Shorts, and Instagram thumbnails and video covers with a full pipeline for concept, casting, scene, render, text overlays, and revisions.

Instructions

Render a click-driving YOUTUBE / Shorts / Instagram THUMBNAIL or video cover through the full production pipeline (concept, casting, scene, render, tweaks, text), not a bare image prompt. Use it for any "thumbnail", "video cover" or MrBeast-style packaging ask INSTEAD of generate_image. Default: 4 variants at 1K (YouTube shows 1280x720 at most), each billed; 2K or 4K only when the user asks. dryRun:true prices it free. The headline overlay is free. WHO IS IN FRAME — never assume or silently substitute a stranger. A framework with a person and no face photo is refused (nothing charged): ask the user once — themselves (a face photo, identity-locked), a generated person (castGenericPerson:true), or a people-free framework. TEXT — default is a CLEAN render with the headline TYPESET over it (free, legible, correctly spelled): pass headline. bakeText:true only on an explicit ask for words painted INTO the image. ALWAYS check the returned postRenderCheck against the image before presenting it. Full rules: get_instructions({topic:'make_thumbnail'}).

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "make_thumbnail"
  }
}

Input Schema

NameRequiredDescriptionDefault
fontNoheadline font: Anton (default) or any Google Fonts family
logoNoa brand logo URL or path to place into the composition
splitNosplit/panel LAYOUT — only when the user asks for one ("split", "before/after", "versus screen"). "X vs Y" as a SCENE stays one unified frame
takesNocamera takes per emotion, 1–4: designed framing / low-angle hero / extreme close-up / wide dutch tilt
topicNothe video's topic — used to pick the hero object when you don't name keyElements
tweakNosurgical pixel-faithful edit of a FINISHED thumbnail (needs sourceImage): kind emotion / background / background_color / rim_light, or any other kind with the edit in words as value
dryRunNotrue = price only, nothing rendered
logo3dNofirst turn the flat logo into a volumetric 3D render (one extra billed image), then composite that
peopleNopeople described in prose instead of by photo (each still gets the chosen expression)
emotionNothe expression on the face (default 'shock') — shock · hype · fear · confusion · determination · smug · charisma · disgust · awe · rage · laugh, or your own phrase
bakeTextNodefault false. true paints the headline INTO the generation — only on an explicit user ask; it leaks garbled text elsewhere in the frame
emotionsNorender one variant per emotion (variants = emotions × takes, max 16)
get_costNotrue: return the price only; nothing runs or bills
headlineNo2–4 word headline. Typeset OVER the finished render by default (free, always legible); newlines split it into stacked lines
locationNoplace, time of day, weather, atmosphere
rimColorNocolored back+hair light — ONLY when the user names one: 'ice-blue' / 'neon-magenta' / 'toxic-lime' / 'amber-gold' / 'pure-white'
variantsNohow many to render (default 4, max 16), each billed; 1 for a single one
frameworkNoconcept framework id (default 'posed_portrait') — before_after · social_ui · three_step · screenshot · posed_portrait · posed_action · specific_day · graphical · landscape · map_aerial · product · adding_text · repetition · size_difference · news_clip · amplified_reality — or your own concept in words
referenceNofields YOU extracted by eye from a reference thumbnail. Extract ALL of: brief (one dense sentence on the concept), subject (pose/action generically, NEVER a specific identity), elements, location, composition, background, split (boolean), split_count, person_count (0-3), emotion (one of the 11 presets or 'other'), emotion_detail (one vivid sentence covering eyes, brows, mouth, head angle). emotion + emotion_detail carry the reference's actual facial performance, which is the single biggest CTR lever on a face; split/split_count reproduce its panel structure. The reference image itself is never sent to the model
backgroundNooverride the default bold saturated colour-field background
faceImagesNoup to 3 face photos (URLs or local paths) — each becomes a locked CHARACTER identity, in order
resolutionNosize; default 1K; 2K or 4K only when asked (each costs more)
sceneBriefNowhat the thumbnail depicts — the concept in one dense sentence, rendered exactly
aspectRatioNo'16:9' (YouTube, default) / '9:16' (Shorts) / '4:5' (Instagram) / '4:3' / '1:1'
bakedUiTextNoshort label for a text-carrying framework (a chat bubble, a DAY N badge, a news lower-third, a map callout) — needs frameworkRequested:true
compositionNooverride the default large-foreground-subject composition
keyElementsNosignature props / effects that make it pop — oversized, flying toward camera
sourceImageNothe finished thumbnail URL a `tweak` edits; tweaks chain, so feed each accepted output into the next
overlayStyleNoheadline style: beast (default), fire, neon-lime, clean-glass, marker, or your own CSS declarations
forceGenerateNorender the 'screenshot' framework anyway (it is normally a real video frame, not a generation)
headlineLinesNoexplicit headline lines (up to 3) — overrides splitting `headline` on newlines
headlinePlaceNobottom (default), top, center, or a 0-1 fraction from the top; never over the face
logoPlacementNooverlay (exact, flat) | in_scene (the real file printed onto a blank panel in the shot; with logo3d, a 3D render of it, checked); auto: in_scene with logo3d
idempotency_keyNochosen once, resent unchanged on retries: the same key never buys twice
restrainedGradeNotrue for a calm / premium / muted look instead of the default punchy poster grade
expected_creditsNothe get_cost price; refused unbilled if the live price differs
castGenericPersonNopass true only after the user has explicitly chosen a generated stranger over their own face
frameworkRequestedNotrue ONLY when the USER named this framework — it is what authorizes a text-carrying framework (social_ui / news_clip / specific_day / map_aerial) to bake its short UI label

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed6 schema fields changedv0.1.532
    • addedInput schema / properties / dryRun
      Added value: +{
      +  "description": "true = price only, nothing rendered",
      +  "type": "boolean"
      +}
    • addedInput schema / properties / expected_credits
      Added value: +{
      +  "description": "the get_cost price; refused unbilled if the live price differs",
      +  "type": "number"
      +}
    • addedInput schema / properties / get_cost
      Added value: +{
      +  "description": "true: return the price only; nothing runs or bills",
      +  "type": "boolean"
      +}
    • addedInput schema / properties / idempotency_key
      Added value: +{
      +  "description": "chosen once, resent unchanged on retries: the same key never buys twice",
      +  "type": "string"
      +}
    • addedInput schema / properties / resolution
      Added value: +{
      +  "description": "size; default 1K; 2K or 4K only when asked (each costs more)",
      +  "enum": [
      +    "1K",
      +    "2K",
      +    "4K"
      +  ],
      +  "type": "string"
      +}
    • changedInput schema / properties / variants / description
      Previous value: -"how many thumbnails to render (default 1, max 16). Each is its own billed render — offer a set of 4 rather than assuming"New value: +"how many to render (default 4, max 16), each billed; 1 for a single one"
  2. Changed1 schema field changedv0.1.519
    • changedInput schema / properties / logoPlacement / description
      Previous value: -"overlay (exact, flat) | in_scene (from the file, checked); auto: in_scene with logo3d"New value: +"overlay (exact, flat) | in_scene (the real file printed onto a blank panel in the shot; with logo3d, a 3D render of it, checked); auto: in_scene with logo3d"
  3. Changed5 schema fields changedv0.1.465
    • addedInput schema / properties / people / items / properties / describe / description
      Added value: +"the person's appearance in English prose, used instead of a face photo"
    • addedInput schema / properties / split / properties / mode / description
      Added value: +"layout: plain, before_after, versus or custom"
    • addedInput schema / properties / split / properties / panels / description
      Added value: +"panel descriptions in order, left to right or top to bottom; sets the panel count (minimum 2)"
    • addedInput schema / properties / tweak / properties / kind / description
      Added value: +"emotion, background, background_color or rim_light, or another kind with the edit in words"
    • addedInput schema / properties / tweak / properties / value / description
      Added value: +"new value for the kind: an emotion, a colour, or the edit in words"
  4. Changed1 schema field changedv0.1.398
    • addedInput schema / properties / logoPlacement
      Added value: +{
      +  "description": "overlay (exact, flat) | in_scene (from the file, checked); auto: in_scene with logo3d",
      +  "enum": [
      +    "auto",
      +    "overlay",
      +    "in_scene"
      +  ],
      +  "type": "string"
      +}
  5. Changed2 schema fields changedv0.1.320
    • changedInput schema / properties / emotion / description
      Previous value: -"the expression on the face (default 'shock') — a preset id or your own phrase"New value: +"the expression on the face (default 'shock') — shock · hype · fear · confusion · determination · smug · charisma · disgust · awe · rage · laugh, or your own phrase"
    • changedInput schema / properties / framework / description
      Previous value: -"concept framework id (default 'posed_portrait'), or your own concept in words"New value: +"concept framework id (default 'posed_portrait') — before_after · social_ui · three_step · screenshot · posed_portrait · posed_action · specific_day · graphical · landscape · map_aerial · product · adding_text · repetition · size_difference · news_clip · amplified_reality — or your own concept in words"
  6. Changed7 schema fields changedv0.1.285
    • changedInput schema / properties / font / description
      Previous value: -"headline font (default Anton). Alternatives incl. Bebas Neue, Oswald, Archivo Black, Montserrat, Inter, Playfair Display"New value: +"headline font: Anton (default) or any Google Fonts family"
    • changedInput schema / properties / framework / description
      Previous value: -"concept framework id (default 'posed_portrait'); see the list in this description / hermoso_capabilities"New value: +"concept framework id (default 'posed_portrait'), or your own concept in words"
    • changedInput schema / properties / headlinePlace / description
      Previous value: -"where the headline sits — never over the face (default 'bottom')"New value: +"bottom (default), top, center, or a 0-1 fraction from the top; never over the face"
    • removedInput schema / properties / headlinePlace / enum
      Removed value: -[
      -  "bottom",
      -  "top",
      -  "center"
      -]
    • changedInput schema / properties / overlayStyle / description
      Previous value: -"headline style: 'beast' (default, white + heavy black stroke) / 'fire' / 'neon-lime' / 'clean-glass' / 'marker'"New value: +"headline style: beast (default), fire, neon-lime, clean-glass, marker, or your own CSS declarations"
    • changedInput schema / properties / tweak / description
      Previous value: -"surgical pixel-faithful edit of a FINISHED thumbnail — needs sourceImage"New value: +"surgical pixel-faithful edit of a FINISHED thumbnail (needs sourceImage): kind emotion / background / background_color / rim_light, or any other kind with the edit in words as value"
    • removedInput schema / properties / tweak / properties / kind / enum
      Removed value: -[
      -  "emotion",
      -  "background",
      -  "background_color",
      -  "rim_light"
      -]
  7. Addedv0.1.161

TDQS

A4.6/5.0
Behavior4/5

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

Annotations only give readOnly=false, destructive=false, openWorld=true, idempotent=false; the description adds substantial context beyond them: per-variant billing, dryRun/get_cost as a free price check, the free headline overlay, a no-charge refusal when a person is framed without a face photo, expected_credits refusal on price drift, and the always-check-postRenderCheck instruction. The only gap is return-shape detail, mitigated by the pointer to get_instructions. No annotation contradiction (idempotency_key is an explicit opt-in, not a contradiction of idempotentHint=false).

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?

It is long and uses heavy capitalization and inline field references, but it is front-loaded with purpose then usage then hard rules, and nearly every sentence carries a decision-relevant constraint. Density is justified by a 38-parameter tool, though some emphasis (ALL-CAPS phrases, repeated free/billed notes) could be tightened.

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 38-parameter, open-world, non-idempotent generation tool with no output schema, the description covers the critical decisions: face/identity handling, text baking vs typesetting, billing and dry-run pricing, resolution defaults, and a pointer to full rules and to the returned postRenderCheck. Nothing an agent needs to invoke it safely and correctly appears missing.

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 100%, so the baseline is 3, but the description earns above baseline by adding cross-parameter semantics the schema cannot: the default of 4 variants at 1K, that variants are each billed, that dryRun prices for free, that headline is typeset by default while bakeText is exception-only, and that castGenericPerson requires prior user consent. It doesn't re-explain most individual fields, which is appropriate given the schema already documents them.

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 opening sentence gives a specific verb and resource ('Render a click-driving YouTube / Shorts / Instagram thumbnail or video cover') and immediately contrasts the scope with a sibling, telling the agent to use this INSTEAD of generate_image. An agent can distinguish it from generate_image and make_* tools without opening any schema.

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

Usage Guidelines5/5

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

It names the alternative (generate_image) and the exact condition that selects this tool ('any thumbnail, video cover or MrBeast-style packaging ask'), plus when NOT to escalate resolution (2K/4K only when asked) and when to set bakeText or castGenericPerson (only on explicit user ask/choice). Guidance covers defaults, exclusions, and escalation paths explicitly.

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

Deploy Server

Other Tools