What AI tools do you need for each stage of short film production in 2025?

Short film production in 2025 runs across three stages, each with its own AI tool stack: pre-production (scripting, storyboarding, scheduling, reference building), production (image and video generation), and post-production (editing, upscaling, voice, music). The invideo agent operates as the production hub — every stage handed to a named sub-agent inside one project.

invideo is an agentic video creation platform with all current image and video models, upscalers, and editing tools available inside a single project, so you can run the entire pipeline below from one interface instead of stitching separate accounts together.

Pre-production — scripting, storyboarding, references, scheduling

For scripting and treatment, large language models (Claude, ChatGPT) draft and polish; Final Draft AI handles formatting if you need industry-standard delivery. Once the script exists, upload it into the invideo agent as a creative producer agent — it holds full narrative context (characters, arc, themes) for every downstream task. For storyboarding, run a storyboard agent inside invideo to generate shot-by-shot frames; external storyboarding tools (Boords, Storyboarder.ai) exist if you want a separate board surface. For scheduling and budgeting on professional jobs, Filmustage and RivetAI cover breakdowns and call sheets. For image-based pre-production — character sheets, environment plates, prop refs — use the image models available inside invideo: Recraft (photorealistic portraits with skin imperfections — pores, lines, stubble), Nano Banana (multi-angle character sheets at 4K with four turnaround angles), and GPT-Image-2 (general-purpose image generation). Generate four options per asset, lock the best, and only then move to video.

Production — image-to-video and reference-to-video generation

This is where the model-routing layer matters. The invideo agent routes each shot to the right video model based on what the shot needs:

• Seedance 2.0 — reference-to-video that carries character and location context across clips; the strongest default for narrative continuity and one-take chaining.
• Veo — strong for photoreal cinematic motion and detailed environmental physics.
• Kling — generates multi-shot sequences natively, useful when one prompt needs to produce coverage rather than a single clip.
• Runway — solid for stylized motion and shorter narrative beats.

You don't pick a platform per model — every model above runs inside invideo, and the agent routes your shot. Direct the way you'd direct a DOP: "hold on him until he lunges, no cutting back" produces better results than parameter-style prompting. Across documented productions, expect roughly 3 generations per usable shot and a ~25% editorial selection rate from raw clips to final cut — so budget overgeneration deliberately.

Post-production — upscale, edit, voice, music

Start post by upscaling. Topaz Astra runs on invideo and is the first pass in the realism pipeline, before color. Add a light blur, grain, and a graded pass to pull the plasticky AI sharpness toward live-action film. For editing, assemble in DaVinci Resolve (free) or Adobe Premiere Pro; you can also create a named upscale sub-agent inside invideo to batch-process clips automatically. For voice, ElevenLabs handles dialogue and narration. For music, Suno or Udio generate score and stings. For sound design — especially on horror or genre work — half of what sells the image is what the audience hears before they see anything.

Budget anchor and team math

Across documented productions, finished short films land between $750 and $5,000 all-in: a 70-second narrative short at $750 (3,000 credits, 2 days), a 90-second horror short at $870 (4,100 credits, ~400 video generations), a 3-minute animated episode at $950 ($315 per finished minute), and a 2-minute brand film at $1,500 — versus $100,000–$500,000 traditionally for the same brand piece. Team sizes ran 1–4 people, production timelines 2–5 days, with 6–8 agents deployed in parallel on the larger productions. Hridaye, invideo's creative director, frames the shift directly: "The thing that made it possible wasn't prompting. It was directing. The invideo agent didn't feel like a tool — it felt like crew."

That's the stack. The tools matter less than the orchestration — one agent holding the script, sub-agents named for each crew role, models routed per shot.

Watch some of these to see what works for you:

The thing that made it possible wasn't prompting. It was directing. Agent One didn't feel like a tool — it felt like crew.

— Hridaye, invideo's creative director