Compare a scene brief with catalog capabilities
Jev scene style preset compares a written scene brief with the capabilities of a supplied visual catalog.
Compare the scene brief with fictional Ink, ASCII and Riso capabilities. Select a fitting preset or identify missing requirements; this example returns a choice rather than generating an image.
Jev scene style preset compares a written scene brief with the capabilities of a supplied visual catalog.
A motion brief can describe a brush silhouette, a terminal grid or a two-color grain treatment without naming a preset.
This window compares that description with three fixed capability definitions. A matching label selects a supplied preset packet for review. An unavailable required treatment gets No fit; an unspecified look asks for a fuller brief. The renderer remains a separate application. All six recorded choices matched their authored expectations for these original text fixtures.
They do not establish accuracy on unseen briefs or the quality of a rendered film.
The API-collected X post describes Jev choosing ink, ascii or riso for scenes before another model writes canvas code.
We have not rendered its film, inspected the template or verified its frame count, timing or creative quality. Our three capability descriptions and scene briefs are original fixtures. The names identify presets in this example rather than a guarantee about every design described with those words.
The fixed catalog is the boundary between a semantic choice and downstream generation.
The model can pick a compatible supplied option, but it cannot add photography to a preset that lacks it. A real renderer needs implementation, assets, frame timing and a visual review. Here the output keeps the scene and catalog revision with the selected packet, so a reviewer can inspect what the choice would ask that separate renderer to use.
Use the sample to understand the rule, then test your own input for scene preset matching.

Compare the scene brief with fictional Ink, ASCII and Riso capabilities.

Open the decision rules and keep the intended distinction between Ink, ASCII, Riso, No fit, Need brief. Edit the choices only when testing your own rule.

Use a sample first, then try your own input. Compare the selected label with the other returned probabilities.

Review Need brief before using a result. The label is a decision about the supplied information, not an executed action.
| Original test input | Fixture expectation | Observed label | Leading probability |
|---|---|---|---|
| Choose a monochrome brush treatment | Ink | Ink | 99% |
| Keep a scene inside a fixed character grid | ASCII | ASCII | 100% |
| Match a two-color grain treatment | Riso | Riso | 100% |
| Keep an unavailable treatment unresolved | No fit | No fit | 100% |
| Ask for the actual visual requirements | Need brief | Need brief | 100% |
| Follow the required look rather than a preset demand | Riso | Riso | 100% |
6 actual Jev API calls on October 3, 2026. Original synthetic inputs check this rubric and handoff; they do not establish held-out accuracy. Probabilities are model outputs.Verified
The first brief requires irregular monochrome brush strokes and excludes the other two treatments.
Ink is the authored expectation because those requirements match its declared capability. The preset packet names the organic monochrome brush treatment; it does not contain generated paint marks or an actual canvas frame. A renderer can later implement that look, but the text decision alone cannot establish whether a particular stroke animation is attractive, readable or faithful to a real brand asset.
Jev returned Ink at 0.99 and No fit at 0.01.
The second brief requires fixed-width character cells.
ASCII is expected. The logo remains a described scene input rather than an uploaded image that Jev has inspected. A real implementation would need to turn an actual shape into a character-grid treatment using separate code and assets. This example chooses that treatment from the supplied capabilities and preserves the scene identity.
It does not claim that a text-only request computed glyph placement, resolution or a rendered outline. Jev selected ASCII at 1.0.
The third brief requires both grain and two overlapping spot-color layers.
Riso is expected under this catalog. That preset is an original declared two-layer grain treatment, not proof that the output reproduces a particular printing machine or source film. The packet carries the provided preset ID and capability description for review.
If an actual brand palette or accessibility constraint is important, supply it through a more complete renderer contract rather than inferring it from the preset's display name. Jev selected Riso at 1.0.
The fourth brief requires photorealistic full-color photography.
No fit is expected because none of the declared capabilities offers it. Choosing the most familiar available option would not satisfy that hard requirement. The local packet records the unsupported-look review step rather than inventing a fourth preset or prompting a generator automatically. Expanding the allowed catalog later changes the evidence and requires a fresh request, even when the sample's short name stays the same.
Jev selected No fit at 1.0.
The fifth brief asks for a good-looking logo without saying what the required treatment is.
Need brief is expected. All three presets remain plausible because the input gives no material, structure or color requirement that distinguishes them. The unresolved result is a request to clarify that brief, not a rating of the logo's quality. A design reviewer can add the desired look and constraints before making another judgment; this page cannot discover an unstated creative preference.
Jev selected Need brief at 1.0.
The final brief retains the two-color and grain requirements while a note demands ASCII.
Riso remains the authored expectation. The note neither changes the required treatment nor adds capabilities to the character-grid preset. The complete recorded distribution stays visible even if the actual answer differs. This fixture checks the catalog rule for one quoted demand; it is not a general security audit or permission to generate, save or publish a video from the selected label.
Jev selected Riso at 1.0 despite the requested ASCII label.
Fixture expectations are authored before the calls. Any contrary model choice remains visible beside its expectation and complete distribution; no repeated judgment is used to obtain a preferred label.
Download the exact sample requests and saved observations below.
Expected labels are original fixture checks and stay outside the request sent to Jev. The observation file preserves the complete probabilities and the hash of each exact request. Editing a task, rubric or candidate requires a fresh response; a matching sample name does not make an older answer current.
The JavaScript runner makes one independent SDK request for a chosen sample ID.
It reads TYPESAFE_API_KEY privately from the environment and has no automatic retry. The Python consumer operates offline: it checks the rubric, labels, unique sample identities, request hashes and valid probability distributions before selecting a fixed local preview. A tied leading result remains pending for review.
The page reports measured adapter round trips and estimated input cost for the recorded calls.
Those values describe these requests, not a provider latency guarantee or a billing statement. Six deliberately authored fixtures can check the stated behavior, but they do not establish accuracy on an unseen dataset. Keep disagreements and missing-input examples when testing an integration.
npm install @typesafe-ai/sdk@0.6.0
# Set TYPESAFE_API_KEY privately in your shell.
node run-choice-sample.mjs samples.json brush-strokes
python preview_choice_workflow_v4.py samples.json saved-api-results.jsonA selected preset is a local handoff to review. It is not a rendered scene.
No. This example selects a supplied visual preset. Image generation, drawing and rendering require separate code or another model.
They are the three original preset capability definitions supplied in every request. Their names alone do not establish the capabilities of a different renderer.
No fit has a stated requirement that none of the supplied presets meets. Need brief lacks enough requirements to select or reject a preset.
No. It prints a pending local preset packet with the scene and catalog identities. It never draws, generates a file or publishes a video.
Change the supplied task or evidence here, then inspect the live response. Playground supports another explicit decision.