Black Forest Labs
Generate storyboard frames on Martini using FLUX.2 — the prompt-fidelity pick for boards where every shot is staged literally and frame composition must match the director's shot list verbatim. FLUX.2 renders explicit camera angles, character positions, and prop placements with literal accuracy, which is exactly what feature-quality pre-viz needs. Pair it with a saved cinematic style prompt, fan into 8-12 frames keyed to the shot list, and feed selected panels into Sora 2 or Kling 3 for animatic.
FLUX.2 reads literal compositional language. Skip "wide shot of the runner" and write: "Wide cinematic shot, runner positioned center-frame at lower third, dawn lighthouse small in the upper-right, mist rising from rocks below, camera at chest height pointing slightly upward, 16:9 aspect." Spell out shot size, subject position, supporting elements, camera height, and aspect ratio — FLUX.2 follows literal positions more reliably than any other Martini image model.
FLUX.2 has no style-reference image input. Save the visual language in a Text node: "Cinematic photography style, anamorphic compression, soft natural light, muted desaturated palette, shallow depth of field, 35mm film grain texture." Wire it into every storyboard node; each node concatenates the style block with its shot-list prompt. Cinematography stays consistent across the board because the prefix is identical.
FLUX.2 Pro's prompt adherence is meaningfully tighter on detailed shot composition. Pin Pro on the establishing shot, the climax frame, and the closing tag — anything where literal staging matters most. Pin base on the medium frames in between. The cost difference matters less than the quality lift on the hero shots.
FLUX.2 cannot accept reference images for character lock — describe the protagonist in detailed terms in every frame prompt: "the same runner, mid-thirties, dark athletic gear with a distinctive red stripe down the sleeve, lean build, distinctive jawline, short dark hair." Detailed descriptions get ~75-85% identity coherence across the board. For tighter character lock, generate the canonical character on Nano Banana 2 first, then describe it from that reference.
On the Martini canvas, place FLUX.2 nodes left-to-right in story order. Each shares the cinematic style block and the character description; each has its own shot-specific compositional prompt. The board reads as a sequence even before frames generate. Iterating one frame is also easy — change one node's prompt without affecting the others.
Each storyboard frame is also a video keyframe. Once approved, route the chosen FLUX.2 frames into Sora 2 or Kling 3 video nodes — image-to-video with the cinematic frame as the starting frame. Add shot-specific motion prompts ("slow camera push," "static frame, runner moves out of frame right"). The animatic ships from the same canvas without re-prompting the video model from scratch.
Establishing shot with literal compositional language. Note the explicit subject position (lower third), supporting element placement (upper right), and camera height — FLUX.2 follows all three.
[Cinematic style prefix] + Frame 1 (establishing): Wide cinematic shot, runner positioned center-frame at lower third, dawn lighthouse small in the upper-right, mist rising from rocks below, camera at chest height pointing slightly upward, the same runner described as: mid-thirties, dark athletic gear with a red stripe down the sleeve, 16:9 aspect.
Story turn with explicit two-shot composition. The "off-screen left" direction guides FLUX.2 to compose with negative space deliberately.
[Style prefix] + Frame 4 (story turn): Medium two-shot, the runner stops at center-frame, pointing toward something off-screen left, a small bird in flight visible in the upper-left third, soft golden side light from camera left, 16:9 aspect.
Tension peak with literal detail ("single bead of sweat tracking down the temple"). FLUX.2 renders this exactly.
[Style prefix] + Frame 8 (tension peak): Tight close-up, runner's face center-frame, eyes locked on something out of frame, single bead of sweat tracking down the temple, dramatic side light from camera right, 35mm film grain prominent, anamorphic compression, 16:9 aspect.
Closing tag with literal compositional rule-of-thirds. The "palette warming to gold" carries the tonal shift cleanly.
[Style prefix] + Frame 12 (closing tag): Wide reverse over the cliff, the runner silhouetted at center-frame against the breaking sun, the path visible stretching ahead in the lower third, palette warming to gold, distant horizon at upper third, 16:9 aspect.
Write literal compositional language ("center-frame at lower third," "camera at chest height"). FLUX.2 follows literal positions more reliably than any other image model in Martini.
Pin Pro tier for hero frames (establishing, climax, tag), base for medium frames. Quality lift on hero shots is meaningful.
For character lock, describe the protagonist in detailed terms each frame. ~75-85% coherence; for tighter lock, generate on Nano Banana 2 first.
Iterate one variable per node, never the whole prompt. FLUX.2 shifts composition on every word change.
Match aspect ratio to the final delivery (16:9 for film/commercial, 21:9 for cinematic widescreen, 9:16 for vertical). Re-cropping panels later breaks the board read.
Save the storyboard canvas as a template after the first board ships. The style block + character description + node graph become the production pipeline.
FLUX.2 returns 1024-2048 wide outputs with literal compositional adherence holding ~90% across 12+ frames from the same style block. Generation time: base 15-30s, Pro 30-60s. Best at literal staging and detailed shot composition; pair with Nano Banana 2 for tight character lock and GPT Image 2 for narrative beat interpretation. Output drops onto the canvas — wire each chosen frame into Sora 2 / Kling 3 for animatic, or chain to Ideogram for any in-frame title cards.
Connect FLUX.2 with other AI models on Martini's infinite canvas. No GPU required — start free.
Get Started FreeMidjourney
Draft a shot list as image nodes laid out left-to-right on the Martini canvas using Midjourney v7 — cinematic frames with editorial mood that read as cohesive storyboard panels for a commercial or short film. Midjourney is the strongest pick when the storyboard needs atmospheric weight and painterly cinematography rather than literal staging. Lock a single style reference, fan into 8-12 frames keyed to script beats, then feed the strongest panels straight into Sora 2 or Kling 3 for animatic motion tests.
View guideOpenAI
Generate storyboard frames on Martini using GPT Image 2 — strong narrative reasoning makes it the right pick when the script reads like a beat-driven story rather than a literal shot list. GPT Image 2 interprets emotional context and composes scenes that carry meaning across panels, which is exactly what directors need when boarding a 30-60 second commercial or short-form narrative. Pair it with a saved style prompt block and feed selected frames into Sora 2 or Kling 3 for the animatic.
View guide