Kleo

Adapt a video request into the film's spec and treatment

kleo_adapt_prompt

Read-only

Step 1 (recommended). Reads the user's request against Kleo's intake and asks for what is missing — the film or the animatic (with this account's prices) and the narration's language (English or Italian) among it; once nothing is, turns it into the SPEC (the request taken apart into the requirements the finished film is checked against: who is in it and exactly how they look, where, what happens in which order, what must be seen, read or said, what must never appear — every item quoting the user) and then the TREATMENT of the film written under it (logline, angle, opening image, acts with their seconds, ending, visual language, pacing, narrator, motifs), listing every decision Kleo took that the user did not ask for. When the user described their film, the film is theirs: their characters as described, their events in their order. Takes the user's reference pictures (https links, or an upload token from kleo_upload_link) and answers with their handles and what they show. Nothing is charged and no GPU is rented. Then show the user what Kleo understood, the logline and the decisions, and pass the "spec" and the "treatment" — unchanged, or corrected as the user asked — to kleo_create_video.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`tone`	No	The tone the user asked for, when they said it.
`music`	No	The user's answer about music, in their words: "no" (in any spelling) for none; "yes", or the kind they want (a mood, a genre, an instrument), for an instrumental track under the narration. Asked every time; pass their answer, never a guess.
`style`	No	The look: "realistic" (filmed, cinematic) or "animation" (a 2D animated film). Read off the request when its words say it; otherwise the tool asks the user for it — pass their answer here.
`author`	No	Who writes the spec and the treatment. "assistant" (default): Kleo hands YOU the two methods and you write them — you are a far stronger writer than Kleo's own planning model, and it costs nothing. "server": Kleo's model writes them (use only if you cannot write JSON yourself).	assistant
`format`	No	16:9 for YouTube/landscape, 9:16 for a Short/TikTok/Reel. Read off the request when its words say it; otherwise the tool asks the user for it — pass their answer here, never a guess.
`prompt`	Yes	The user's request in their own words.
`product`	No	The user's answer to the question film or animatic, which the tool asks with this account's prices: "film" (every shot a generated clip, for accounts that have bought a pack) or "animatic" (the drawn frames with camera moves, 5 credits up to 20 seconds, 10 up to 60 seconds). Never a guess.
`audience`	No	Who the video is for, when the user said it.
`language`	No	The user's answer to the question about the narration's language, in their words ("en", "English", "italiano"…): Kleo narrates in English or Italian, and "whatever" / "no preference" means English. Asked every time the request does not say it: pass their answer, never a guess, and never the language they write to you in.
`must_keep`	No	What the user said must appear (names, numbers, places, a message) or must not, when they answered that question.
`subtitles`	No	The user's answer about burned-in subtitles: "yes" for thin cinema subtitles in the video, "no" for none (an .srt file is delivered either way). Asked every time; pass their answer, never a guess.
`ai_upscale`	No	The user's answer to the AI upscale question, in their words ("yes" / "no"): asked for a film only, with its price in extra credits. Only a clear yes buys it; "no", "whatever" or no answer means the classic 4K 60 fps finish. Never a guess.
`duration_s`	No	Length in seconds. Read off the request when its words say it; otherwise the tool asks the user for it — pass their answer here, never a guess.
`references`	No	The pictures the user gave for this film, each with ONE of: url (a public https link), upload (the token of a kleo_upload_link link they uploaded through), handle (kref_… returned earlier); plus role and name when the user said what it shows. Kleo draws the characters, objects and places FROM these pictures.

Schema Changelog

Changes observed during successful MCP inspections.

Changed1 schema field changedSep 27, 2026

changedInput schema / properties / product / description

Previous value: -"The user's answer to the question film or animatic, which the tool asks with this account's prices: \"film\" (every shot a generated clip, for accounts that have bought a pack) or \"animatic\" (the drawn frames with camera moves, 5 credits flat, up to 60 seconds). Never a guess."New value: +"The user's answer to the question film or animatic, which the tool asks with this account's prices: \"film\" (every shot a generated clip, for accounts that have bought a pack) or \"animatic\" (the drawn frames with camera moves, 5 credits up to 20 seconds, 10 up to 60 seconds). Never a guess."

Changed1 schema field changedSep 25, 2026

addedInput schema / properties / ai_upscale

Added value: +{
+  "description": "The user's answer to the AI upscale question, in their words (\"yes\" / \"no\"): asked for a film only, with its price in extra credits. Only a clear yes buys it; \"no\", \"whatever\" or no answer means the classic 4K 60 fps finish. Never a guess.",
+  "maxLength": 80,
+  "type": "string"
+}

Changed4 schema fields changedSep 25, 2026

changedInput schema / properties / language / description

Previous value: -"Language of the narration; detected from the request when omitted."New value: +"The user's answer to the question about the narration's language, in their words (\"en\", \"English\", \"italiano\"…): Kleo narrates in English or Italian, and \"whatever\" / \"no preference\" means English. Asked every time the request does not say it: pass their answer, never a guess, and never the language they write to you in."

removedInput schema / properties / language / enum
```
Removed value: -[
-  "en",
-  "it"
-]
```
addedInput schema / properties / language / maxLength
```
Added value: +40
```

addedInput schema / properties / product

Added value: +{
+  "description": "The user's answer to the question film or animatic, which the tool asks with this account's prices: \"film\" (every shot a generated clip, for accounts that have bought a pack) or \"animatic\" (the drawn frames with camera moves, 5 credits flat, up to 60 seconds). Never a guess.",
+  "enum": [
+    "film",
+    "animatic"
+  ],
+  "type": "string"
+}

Changed2 schema fields changedSep 24, 2026

changedInput schema / properties / author / description

Previous value: -"Who writes the treatment. \"assistant\" (default): Kleo hands YOU the producer's method and you write it — you are a far stronger writer than Kleo's own planning model, and it costs nothing. \"server\": Kleo's model writes it (use only if you cannot write JSON yourself)."New value: +"Who writes the spec and the treatment. \"assistant\" (default): Kleo hands YOU the two methods and you write them — you are a far stronger writer than Kleo's own planning model, and it costs nothing. \"server\": Kleo's model writes them (use only if you cannot write JSON yourself)."

addedInput schema / properties / references

Added value: +{
+  "description": "The pictures the user gave for this film, each with ONE of: url (a public https link), upload (the token of a kleo_upload_link link they uploaded through), handle (kref_… returned earlier); plus role and name when the user said what it shows. Kleo draws the characters, objects and places FROM these pictures.",
+  "items": {
+    "properties": {
+      "handle": {
+        "description": "A handle (kref_…) this tool or the upload page returned earlier.",
+        "pattern": "^kref_[0-9a-f]{8}$",
+        "type": "string"
+      },
+      "name": {
+        "description": "The character's name, when the picture is of a character (\"Mara\", \"my dog Pepe\").",
+        "maxLength": 60,
+        "type": "string"
+      },
+      "note": {
+        "description": "What the user said about the picture, in their words.",
+        "maxLength": 300,
+        "type": "string"
+      },
+      "role": {
+        "description": "What the picture is for: character (draw this person/animal/creature), object, place, or style (imitate the look).",
+        "enum": [
+          "character",
+          "object",
+          "place",
+          "style"
+        ],
+        "type": "string"
+      },
+      "upload": {
+        "description": "The token of a kleo_upload_link link the user uploaded pictures through: every picture uploaded there is taken.",
+        "maxLength": 400,
+        "type": "string"
+      },
+      "url": {
+        "description": "A public https:// link to the picture (PNG, JPEG or WebP, up to 12 MB).",
+        "maxLength": 2000,
+        "type": "string"
+      }
+    },
+    "type": "object"
+  },
+  "maxItems": 8,
+  "type": "array"
+}

First observedSep 22, 2026

TDQS

A4/5.0

Behavior4/5

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

Discloses cost ('Nothing is charged and no GPU is rented'), that it interrogates the user for missing intake fields, and that it returns reference handles naming what each picture shows. This adds real context beyond the readOnlyHint/openWorldHint annotations, which already cover the safety profile.

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

Conciseness3/5

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

Front-loads 'Step 1 (recommended)' and the core action, but the body is a sprawling multi-clause sentence with nested parentheticals that is hard to parse. Most content earns its place, yet it is materially longer and less structured than it needs to be.

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 tool with no output schema, the description covers the flow, the two returned artifacts, reference handling, and the zero-cost guarantee. It is complete enough to invoke correctly, though a clearer statement of the exact return payload shape would close the remaining gap.

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

Parameters3/5

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

Schema description coverage is 100% and each parameter (product, author, language, references, etc.) is fully documented in the schema itself. The description reinforces the 'pass their answer, never a guess' convention and adds conceptual framing for product/author, but offers little parameter-level detail the schema doesn't already provide.

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?

States a specific verb and pair of artifacts (turn the request into the SPEC and the TREATMENT) and defines exactly what each contains. It self-identifies as 'Step 1' and routes onward, so an agent can place it in the pipeline vis-à-vis kleo_create_video 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 Guidelines4/5

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

Clearly establishes this as the first, recommended step and names kleo_create_video as the destination for the 'spec' and 'treatment' it produces. It stops short of stating when NOT to use it or naming an alternative path, but the sequencing guidance is explicit and actionable.

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

All Kleo tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Adapt a video request into the film's spec and treatment

Input Schema

Schema Changelog

TDQS

Glama MCP Gateway

Resources