Tired Things Universe Episode Maker · publiczny skill Open Skill
Tired Things Universe Episode Maker
Tired Things Universe Episode Maker to publiczny Open Skill w Digen. Otwórz go, by zobaczyć, co tworzy ten workflow.
ShortDrama
2 2 1
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4177049/2026-09-25/zgbhpupz.jpg
# Tired Things Universe Episode Maker ## When to use - The user wants a new episode: an everyday object (soap, toothbrush, coffee mug, remote, sponge, toilet paper…) that is "tired" and complains about its daily life. - The user wants to advance or redo one step of an episode (rewrite the script, new character, redo the voiceover, next clip, final mix). ## Reference files - `references/style-guide.md`: The established series style (3D cartoon look, tired expressions, "Tired Things Universe" branding, POV caption, subscribe end card) and the script tone. Read it before generating any script or character. ## Preset assets - `tired_things_reference`: The user's original toothbrush episode video. Use `recognize_video` on it (or read the style guide) to match the established look. ## Workflow (step-by-step — pause after each stage for user approval) First call `write_todos` to list the stages, updating the list as you progress. Each stage produces an asset and pauses for the user to approve or request changes before the next stage. ### Stage 1 — Script Ask which everyday object this episode is about (or take the user's choice). Write a ~10s script where the object complains about its daily life in the series tone (wholesome, melancholic, absurdly relatable — see style-guide). Include the POV caption line. Save it with `create_document(content=..., name="<Object> script", entity_type="scripts")`. Show the script and ask for approval. ### Stage 2 — Character still Generate the character image in the established style. Use `generate_image(model='t2i', orientation='portrait', prompt=<detailed style from style-guide + the object + tired expression + "Tired Things Universe" towel>, name="<Object> character", entity_type="characters")`. If the user provides a reference image, use `model='i2i'` with `input_urls=[that image]` instead. Show it and get approval. ### Stage 3 — Voiceover Generate the object's voiceover from the script. Use `generate_tts(model='tts.design', voice_instruction=<a tired, weary, slightly dramatic voice>, text=<script lines>, name="<Object> voiceover", entity_type="voiceovers")`. Get approval. ### Stage 4 — Talking clip Generate the animated clip of the object talking/complaining, driven by the voiceover. Use `generate_video(model='ai2v', input_urls=[character still], audio_url=<voiceover asset>, orientation='portrait', duration=10, name="<Object> clip", entity_type="storyboard_videos")`. Get approval. ### Stage 5 — Final 10s short Assemble the final short: the talking clip plus the "Please Subscribe!" end card, with the POV caption overlay. Generate the end card image in the Stage 2 style if needed, then `mix_clips` to stitch the clip + end card into the final vertical short. Use `mix_clips(clips=[...], orientation='portrait', name="<Object> final short", entity_type="final_videos")`. Optionally add background music with `generate_music`. ## Notes - Keep the series consistent: same 3D cartoon style, tired expressions, "Tired Things Universe" branding, POV caption, subscribe end card. - Intermediate stage products will be visible on the canvas — that's expected for a step-by-step workflow. - Report failures honestly; never fabricate URLs. - After the final video, call `set_guidance` to suggest next actions (new episode, redo a step, tweak the style).