Studio productions

The studio production tools create and change films that open in the Nodaro Studio editor at studio.nodaro.ai. This page lists each tool's parameters. Studio productions explains the loop, the words to use with the user, the edit operations and the spending rules; read it first.

Studio productions are a Nodaro Cloud feature. On a deployment without them, every tool in this family answers not_available. Call list_studio_productions to check before you offer the feature.

Tools, permissions and prices

The tools that spend credits, and share_studio_production, which changes who can see the work, carry a confirmation mark in their definition. The spending tools accept client_request_id, a retry token of 8 to 128 characters: send the same value again after a timeout, and nothing is started or charged twice.

Plan and read

get_studio_production_skill

Returns the guide an assistant reads before it works on a production. It is generated from the live deployment.

validate_studio_plan

Checks an authored plan, free, and saves nothing. It also resolves every cast name against your own characters, locations, objects and creatures.

Returns: valid, errors (each naming its field), warnings and a summary that says how many cast names matched your library. Fix the errors and call again until valid is true.

list_studio_productions

Lists your productions, newest first: id, name, version, thumbnail, whether it is shared, and the number of scenes. Archived productions are hidden, as on the dashboard.

get_studio_production

Returns one production: its film look, cast, folders, cuts, bin, what is running, and its scenes in timeline order. With workflows:write, it first lands every job that finished since the last read, so this read is how a finished generation reaches its scene.

Returns: the production, and a pending block that names what is still running. Address a result by its key, the job id or the URL, never by its position.

plan_studio_export

Returns the ordered steps that would export the production, with a credit estimate. It starts nothing and charges nothing: show it to the user before they accept.

Create and change

create_studio_production

Creates a production in your own Studio project. It appears on your dashboard at studio.nodaro.ai at once. With a plan, every scene, cast binding and film look lands with it; without one, the production is empty. Validate the plan first.

import_studio_production

Adds a plan's scenes to an existing production, and enrolls the new cast members. It never renames the production or changes its brief or its look.

edit_studio_production

Changes a production with a batch of operations, such as renaming a scene, reordering the timeline, selecting a take, setting the shots inside a scene's motion, enrolling a cast member or emptying the bin. The operations and their arguments are served by get_studio_production_skill with part: "operating".

Returns: receipts, one past-tense line per operation. The batch is applied as one step: if one operation is refused, the answer names its index and nothing is written. A batch composed against a slightly older version still applies, and rebased: true says so.

share_studio_production

Publishes the production to a share link, or makes it private again. Anyone with the link can read a shared production, so ask the user first. This tool is the only way sharing changes.

clone_studio_production

Copies a production, yours or one shared with you, into your Studio project. The copy starts private and carries the graph and every landed result. Use it before a risky round of edits.

Generate

A generation returns job ids at once. The finished result joins its scene only on the next get_studio_production; get_job and wait_for_job report the status and land nothing.

describe_studio_production

Hands a brief to the Director, which writes scenes, cast and looks into the production. Use it when the user has a story rather than a plan.

generate_studio_still

Generates candidates for a scene's frame, from what the scene already says: its prompt, references, cast and direction. Generating again adds takes; it never replaces one.

generate_studio_keyframe

Generates one candidate for a planned frame, a frame the user reviews and accepts, which is different from a scene's frame. A derived frame needs an accepted parent. There is no quote for this tool, and dry_run: true is refused.

Landing the candidate neither accepts it nor starts another frame. Accept it with edit_studio_production after review. This tool needs a deployment that supports planned frames.

generate_studio_clip

Generates a scene's motion from its frame, its start and end frames and its direction.

To retake a linked take exactly, first call with retake_result_key and dry_run: true, and review the price and the inputHash. Then submit the same call with expected_input_hash and a new client_request_id, without dry_run. The retake uses the take's original settings and endpoint images, even if the plan changed since. Deployments advertise this with the operations.retakeLinkedClips capability.

new_studio_shot_from_frame

Takes a frame from a scene's current motion and puts it to work. The call waits for the extraction and answers with the updated production.

voice_studio_shot

Speaks a line over a scene, as a voiceover. The call waits for the speech and answers with the updated production.

revoice_studio_clip

Replaces the voices inside a scene's current motion: the dialogue is performed again and mixed back over the same picture.

score_studio_production

Writes one soundtrack for the whole production from a description of the music.