agentsclimarketplace

Codex dreamina video studio

Skill baitalab/codex-dreamina-video-studio

Give Codex a story. Get an editable AI-video package powered by Dreamina: shots, prompts, generation, editing, SRT, stems, and QA.

Install
npx -y skills add baitalab/codex-dreamina-video-studio

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Turn a story, novel excerpt, treatment, or screenplay into a complete AI-video production and editable post-production package. Use when Codex needs to adapt narrative, design characters and locations, write a timed shooting script, generate storyboard and image-to-video prompts, manage generation budgets and approval gates, assemble clips, create SRT and role-separated dialogue text, export silent picture and aligned music/dialogue/ambience/SFX stems, or perform audiovisual QA.

SKILL.md

3.6 KB, as published. Nobody here has run it

Story to Video Studio

Produce a reusable editing package, not only a flattened MP4.

Workflow

  1. Initialize a project:
python scripts/init_project.py "project-folder" "title"

  1. Read references/narrative-workflow.md and write:

    • adaptation brief
    • character and location bible
    • timed screenplay
    • shot manifest
    • narration/dialogue plan

  2. Read references/visual-generation.md. Create stable character references before shot generation. Write both still-image prompts and motion prompts for every shot.

  3. Estimate cost with scripts/estimate_budget.py. Stop at the generation approval gate unless the user has explicitly authorized autonomous generation within a stated budget.

  4. For Dreamina/即梦, read references/dreamina-cli.md and use scripts/dreamina_provider.py. Preserve prompts, task IDs, models, costs, seeds, sources, and licenses.

  5. Assemble the picture according to references/editing-audio.md. Avoid automatic transitions by default; use hard cuts unless a transition has a narrative function.

  6. Export dialogue assets:

python scripts/export_dialogue.py voice_plan.json delivery-folder

  1. Export an editing package:

    • silent, text-free picture master
    • full mix
    • music stem
    • original ambience stem
    • designed SFX/rhythm stem
    • narration/dialogue stem
    • optional one stem per music cue and speaker
    • full SRT, role-separated SRT, and role-separated plain text

  2. Run:

python scripts/qa_package.py delivery-folder

Resolve duration mismatches, one-frame flashes, clipping, squeals, low dialogue, burned text, missing provenance, and incomplete deliverables before handoff.

Approval gates

  • Gate A: approve adaptation direction, target duration, style, aspect ratio, and budget.
  • Gate B: approve screenplay, shot manifest, character/location bible, and generation estimate.
  • Gate C: approve one representative generated shot before bulk generation.
  • Gate D: approve rough cut before expensive regeneration or final mastering.

If the user explicitly says to continue autonomously within a budget, proceed through gates while logging decisions and never exceed the limit.

Dreamina safety

  • Discover the executable from DREAMINA_CLI or PATH; never hardcode a user path.
  • Keep OAuth state in the official CLI store; never copy login files into a project or the Skill.
  • Dry-run generation commands unless the user has authorized credit spending.
  • Record credit balance before and after a generation batch.
  • Never print or commit cookies, tokens, device codes, user codes, account identifiers, or home-directory paths.

Delivery rules

  • Align all audio stems from 00:00 to the same exact duration.
  • Export WAV at 48 kHz, 24-bit, stereo unless the user requests otherwise.
  • Export picture without subtitles or burned text when editable delivery is requested.
  • Keep original generated clip audio on a separate stem.
  • Keep each speaker separable.
  • Preserve a machine-readable manifest and a reader-friendly production summary.
  • Give the user the containing folder link, not individual file links.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.