FXI Studio

Cancels one of your generations that is still running and refunds its credits in full. If it finishes right before the cancel lands, you keep the result and nothing is refunded. Calling this on an already-terminal generation is safe and returns its current state (no double refund). Prefer lifecycle_status over the legacy status word. Does not claw back upstream compute already spent.

cancel_generation

Checks an upload started with create_upload_url: the file arrived, is within the size limit and really is the declared type, and reports its media type, size, dimensions and duration. Returns the asset_url to pass to generation tools. A file that fails the check is deleted. Generation tools run the same check themselves, so this call is optional.

confirm_upload

Returns how many credits are left on the signed-in FXI Studio account and the name of its plan.

get_credits

Returns the status, per-clip progress, credits used and clip URLs for a frame_chain or compose_scene task. Status is pending, running, succeeded, failed or cancelled.

get_task

Generates one short video clip from a written brief, optionally using up to 4 of your saved characters as references and one of your saved or uploaded images as the opening frame. Spends credits. Returns a task id; check it with get_task. For full control over a single shot, use generate_video instead.

compose_scene

Generates one video clip from a prompt and saves it to your FXI library. Optional start and end frames, reference images, a reference or source video (for example your own clip to edit with Gemini Omni) and reference audio. Bring your own files with create_upload_url or upload_from_url first. Spends credits. Returns a generation id; check with get_generation (usually 1-3 min; prefer lifecycle_status).

generate_video

Generates one image from a prompt, with an optional reference image and aspect ratio, and saves it to your FXI library. Spends credits. Returns the image URL when ready (usually within 25 seconds), or a generation id to check with get_generation (prefer its lifecycle_status field over the legacy status field). Uses the fast default model unless the user picks another one.

generate_image

Generates 2 to 8 video clips in order, where each clip starts from the last frame of the previous clip so motion carries across shots. The first clip starts from your image. Spends credits per clip; a clip that cannot start is not charged. Returns a task id; check progress and clip URLs with get_task. Returns separate clips, not one joined file.

frame_chain

Returns one of your saved generations by id: media URL, prompt, type, model, lifecycle_status and credits used. Prefer lifecycle_status (queued/running/succeeded/failed/cancelled/expired - took too long, credits auto-returned); status is the older 3-value word for legacy clients. created_at is UTC/ISO 8601. wait_seconds (max 25s) long-polls instead of an immediate running; cancel_generation stops a running one.

get_generation

Lists available image/video models, grouped by role, with output type, estimated credit cost, price tier, mode/latency, and a one-line description. Each "model" value passes to generate_image or generate_video. Filter by role or output_type, or sort "cost" for cheapest first. Video estimates assume a 5s clip; pass estimate_only to generate_image/generate_video for an exact quote. Each entry also has speed_tier (fast, balanced or quality), credits, typical_seconds when known, and best_for. Use these to offer a faster or higher-detail option.

list_models