# Shot Card Desk > Write shot lists that earn their place. Five lanes over one scene idea, plus a free > browser-side scorer that reads every card against Walter Murch Rule of Six and reports > what each shot is worth and which one is costing the most. URL: https://shot-card-desk.skillsafe.ai/ API: https://shot-card-desk.skillsafe.ai/api.html Derived from smixs/visual-skills (https://github.com/smixs/visual-skills) by Serge Shima, licensed CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/): its dramaturgy layer - scene formula, details law, three-jobs rule, Murch Rule of Six, blocking and staging, the fourteen-field shot card, the rhythm ladder, the one-anchor principle - and its universal rules for every video model. Not affiliated with or endorsed by Serge Shima, Walter Murch, or any director named in the reference library. ## The scoring model Everything below runs in the browser, on text you paste, with no account and no credits. Murch Rule of Six is a weighting, and each item outweighs everything beneath it combined. A shot that respects screen direction and spatial geometry but misses the emotion scores 9, not 66 - which is why the output is a number rather than a list of ticks. - **Emotion (51)** - Does the cut honour what the viewer feels now, and what they should feel next? Earned by: an emotion named for the shot, and a body carrying it. - **Story (23)** - Does the cut advance story or reveal character? Earned by: a beat - what changes here - and an action that is a real physical event. - **Rhythm (10)** - Does the cut fall on a musical beat of the scene? Earned by: a duration and a cut type. - **Eye-trace (7)** - Where is the gaze at the moment of the cut, and does the next shot receive it? Earned by: an eye-trace note saying where the eye lands in the first 0.3s. - **2D plane (5)** - Does the cut respect the axis of screen direction? Earned by: a composition note - where in the frame, which edge, which direction. - **3D space (4)** - Does the cut respect the geometry of the real location? Earned by: a frame size and a camera whose move has a stated reason. The score is the mean across cards. The panel also names the single criterion losing the most points across the whole list, because that is the one worth fixing first. ## The fourteen-field shot card A card is a `Shot NN` heading with one labelled line per field beneath it. Each field is priced by what it feeds: - `Shot ID` - no weight; direction for the shoot - `Beat` - feeds Story (23) - `Emotion` - feeds Emotion (51) - `Frame` - feeds 3D space (4) - `Composition` - feeds 2D plane (5) - `Camera` - feeds 3D space (4) - `Movement reason` - feeds 3D space (4) - `Action` - feeds Story (23) - `Eye trace` - feeds Eye-trace (7) - `Duration` - feeds Rhythm (10) - `Cut type` - feeds Rhythm (10) - `Sound` - feeds Emotion (51) - `Light / color` - feeds Emotion (51) - `Production note` (optional) - no weight; direction for the shoot ## Words that do not render Each is a placeholder where a physical fact should be, and each has a replacement: - "cinematic" -> name the lens, the light source and its direction - "professional" -> name what makes it look considered - the lighting ratio, the lens - "high quality" -> delete it; it describes nothing the model can render - "masterpiece" -> delete it; tag-spam is an image-model habit and video models punish it - "stunning" -> name the thing that would stun someone - the specific sight - "epic" -> name the scale concretely - how far, how high, how many - "amazing" -> delete it and describe what is actually in frame - "beautiful lighting" -> name the source, direction and colour temperature - "dynamic camera" -> name one move and say what changed to motivate it - "intense moment" -> name the physical pressure - what tightens, what is at stake - "powerful scene" -> name the power relation in the staging - who stands, who sits - "dramatic close-up" -> say whose face, what it is doing, and what the eye lands on - "establishing wide shot" -> say what the geography needs to make readable, and why now - "4k" -> delete it; resolution is a render setting, not direction - "photorealistic" -> name the film stock, lens and light instead An emotion named without a body is the same failure at a larger scale. "He is scared" renders nothing; a locked jaw, a held breath and curled fingers render. Two to four observable cues per emotional transition - below two the model guesses, above four it reads as overacting. ## The three-detail rule Every shot owns at least one of each. A shot with none is filler; a shot with one is thin. - environmental pressure - physical micro-action - sound anchor or visual motif ## Rhythm ladders Drama rhythm is stepped, not uniform. Always at least one pause before the biggest cut. - `slow_burn` - Slow-burn drama: 4, 4, 3, 2, 1, pause, 2 - `commercial` - Commercial product arc: 3, 2, 1.5, 1, 0.5, 2 - `anxiety` - Anxiety build: 2, 1, 1, 0.5, 0.5, 0.3, pause, 1 - `impact` - Impact scene: pause, 0.2, 2 ## Director references One dominant reference per scene, translated into camera, light, rhythm and blocking. Never stack three - they average into nothing. - `fincher` - Fincher: precise motivated camera, no handheld drift, cold palette - `kurosawa` - Kurosawa: weather as emotional pressure - rain, wind, heat - `spielberg` - Spielberg: readable staging, clear geography, emotional wide shots - `wright` - Edgar Wright: sound-driven montage, match cuts on sound events - `glazer` - Jonathan Glazer: a music-video visual idea translated into drama - `wongkarwai` - Wong Kar-wai: longing, reflections, slow emotional drift - `safdie` - Safdie: anxiety, handheld pressure, overlapping sound - `anderson` - Wes Anderson: symmetry, controlled blocking, colour blocks ## The five lanes ### `treatment` - Treatment and anchors Commit to five anchors and one director reference before any shot exists. A short piece that has not chosen its emotion, its motif and its final image will discover them halfway through the shot list, and the first half will not match. Sections: `Summary`, `Five Anchors`, `Treatment`, `Reference`, `Next Step`. Fields: `idea`*, `seconds`*, `director`, `emotion`, `audience`. ### `cards` - Shot cards Fourteen fields per shot, because a missing field is missing direction. The scorer below reads them against Murch's Rule of Six and tells you which shot is weakest and by how much - before you generate anything. Sections: `Summary`, `Shot Cards`, `Card Table`, `Score Check`, `Next Step`. Fields: `idea`*, `anchors`, `seconds`*, `shots`, `draft`, `director`. ### `rhythm` - Rhythm ladder Drama rhythm is stepped, not uniform. Four named ladders, each of which puts at least one pause before the biggest cut - because acceleration needs something to accelerate away from. Sections: `Summary`, `Ladder`, `Rhythm Table`, `Cut Notes`, `Next Step`. Fields: `cards`*, `ladder`, `seconds`*, `climax`. ### `prompt` - Model prompt Turn one card into a prompt a video model will actually render. Full cinematic sentences, one primary camera move, a stated lens, an identity block repeated in every clip - and none of the words that do not render. Sections: `Summary`, `Prompt Plan`, `Video Prompt`, `Language Check`, `Next Step`. Fields: `card`*, `model`*, `lens`, `identity`, `draft`. ### `score` - Score a shot list Paste a shot list you already have. Every shot is scored against the Rule of Six, the weakest one is named with what it is costing, and the lane returns a repaired list rather than a column of complaints. Sections: `Summary`, `Verdict`, `Score Table`, `Findings`, `Repaired Cards`, `Next Step`. Fields: `cards`*, `seconds`, `ladder`, `focus`. ## What this app does not do It writes plans, cards and prompts. It does not generate video or images, it calls no generation API on your behalf, and it edits nothing. The prompts are yours to paste into Seedance, Kling, Veo or anything else.