Skip to content
AI Primer
workflow

Magnific publishes GPT-6 Astra motion-graphics guide for its ChatGPT MCP

Magnific published prompts and iterative directions for making motion-design videos through its MCP in ChatGPT. The guide covers references, character actions, camera changes, object-triggered graphics, and revision prompts.

5 min read
Magnific publishes GPT-6 Astra motion-graphics guide for its ChatGPT MCP
Magnific publishes GPT-6 Astra motion-graphics guide for its ChatGPT MCP

TL;DR

  • Magnific’s city-film prompt treats a 30-second motion piece as a production brief, with reference boundaries, art direction, shot timings, transition language and an audio direction in one request.
  • Graphics get physical triggers rather than floating over footage: Magnific’s packaging-film example makes an orange and produce mesh leave separate printed traces on a paper bag.
  • Character continuity is tracked through a single prop and its state after each action, as Magnific’s courier film follows one audio cartridge from hand to slot.
  • The guide turns the first render into a revision pass, with Magnific’s basketball example specifying camera changes, a completed score and the final aerial move.

An orange signal causes slats, grids and a final moiré facade to appear across a Mediterranean street in Magnific’s 12-cut city film. A paper bag inherits the texture of fruit and plastic mesh in the packaging prompt, while the basketball revision requires the ball to be visibly below the net by 24.8 seconds.

Twelve-cut city film

Magnific begins with a reference clip, then limits what it may contribute: editing energy, mixed-media compositing and the relationship between real footage and graphics. The requested locations, forms, story and compositions must be new.

Its 30-second example breaks the request into five production layers:

  • Art direction: a sun-bleached industrial district, acid orange, near-black cutouts, metallic filaments and moiré grids.
  • Format: native 16:9, 12 distinct shots, cuts roughly every two to three seconds.
  • Story engine: one orange zigzag travels from a wall junction box to a rooftop antenna, producing each disruption along the way.
  • Edit grammar: split screens, scan distortion, paper-cut geometry, snap orbits and two-to-four-frame graphic transitions.
  • Sound and exclusions: a 150 BPM instrumental breakbeat, no speech, logos, watermarks or readable text.

The unusually useful bit is the path. The signal supplies a single direction of travel, so the motion-design events have a shared cause and a final destination.

Cause-and-effect graphics

The packaging-film prompt states one magical rule: when a material touches the living paper, a flat printed echo travels outward from that contact point. The original material remains intact.

That rule dictates the sequence:

  1. An intact orange rolls into a blank, flattened bag and wakes it with a crease ripple.
  2. The bag unfolds under the orange, which leaves dimpled orange contact prints as it moves.
  3. A strip of produce mesh touches the paper and transfers a black lattice, while the physical mesh relaxes back beside the crate.
  4. The two prints lock into a composition on the paper before its existing gussets fold around the same orange.

Magnific also calls for a visible origin and destination for every transition, keeping the orange, bag, mesh and market geography present through the transformation.

Character and prop continuity

For a 15-second 3D mini-film, Magnific identifies the courier’s appearance in detail, assigns her one translucent orange audio cartridge, and tells Astra to retain both across every cut.

The object becomes the continuity check: she holds it while jumping between a moving fader cap and a drum pad, carries it through a chute, inserts it into a matching slot, then has an empty hand. Her delivery activates the factory only after the insertion, when a waveform propagates through its machinery.

The same prompt pairs those state changes with graphic accents tied to contact, including a halftone impact wedge on landing, hand-drawn spark strokes during the slide and a comic split screen that keeps the cartridge visible.

The revision pass

Magnific frames the next conversation turn around concrete editorial changes: closer looks and exchanged glances for tension, faster action for pace, fresh camera angles and a defined payoff.

Its basketball sequence uses those notes as timed constraints:

  • Eye close-ups, a single controlled bounce and a top-down two-shot establish the confrontation before the drive.
  • One right-to-left crossover launches the only burst, which lasts under 0.7 seconds and advances toward the same hoop.
  • At 24.2 seconds, the ball must descend from above, pass through the center of the rim and be visibly below the net by 24.8 seconds in one unbroken take.
  • A crane move begins at 25.2 seconds and reaches a high-oblique view of the court by 29 seconds.

That degree of timing turns the revision prompt into an edit brief, not a general request for extra energy.

Connector and credits

Magnific’s guide finale says the workflow starts with an idea and references, then keeps camera, pacing and graphic refinements in the same ChatGPT conversation. Its MCP landing page lists image, video, audio, vector and 3D generation alongside upscaling, relighting, layer edits and resizing.

According to Magnific’s MCP documentation, the connector is available on paid plans and every MCP action consumes credits, including models that may have unlimited generations elsewhere in a plan. The ChatGPT MCP overview describes the same continuity across references, campaign variations, approved-asset upscaling and image-to-video work.

Further reading

Discussion across the web

Where this story is being discussed, in original context.

On X· 1 thread
Connector and credits1 post
Share on X