How can I see the credit cost before generating an AI video?

Ask the invideo agent for a generation breakdown before you hit render — it returns the planned image count, video clip count, and credit estimate so you know the spend up front. Translate credits to dollars using your plan rate (about $0.21/credit on the generative plan), and run a cheap image probe before spending video credits.

Start every shot or batch by asking the invideo agent for an estimated generation breakdown — "before we generate, give me the planned reference sheets, image count, video clip count, and total credits." The agent returns an itemized plan you can approve, edit, or scale back before any credits are spent. One creative director described it this way: "It also gives me an estimated generation breakdown… which is great because I know what I'm roughly going to spend now before I actually spend it."

Convert credits to dollars with your plan's per-credit rate. On invideo's generative plan that works out to roughly $0.21 per credit, so a 155-credit campaign lands near $33, a 500-credit UGC ad lands near $125, and a 2,100-credit montage lands near $530. Documented productions came in at $67 (270 credits, real estate B-roll), $73–$195 per ad (UGC and product films), and $150 for a 40-still/30-clip editorial campaign (630 credits) — so once you have the agent's credit estimate, multiply by your rate to get the USD figure before you commit.

Probe before you batch. Generate one still image of the shot first — image credits are roughly 65% cheaper than video on invideo, and iterating on a $1–$2 image until the framing is locked beats burning a $5–$15 video render on a composition you'll discard. Only spend video credits on locked frames. For longer runs, generate a single representative "probe shot" (character, garment, set, pose, all in one frame) and confirm the system holds before letting the agent batch the rest.

Use smaller batches to keep the estimate tight. Generating in groups of 5 clips gives you a checkpoint to re-estimate spend between batches — if a batch's yield is low (one production saw 39 clips generated for 10 used, a 26% yield), you adjust the next batch's count rather than committing all 15+ clips at once. Across documented runs, expect to discard a meaningful share of generations (one production rejected ~85% of clips); price that rejection in when you read the agent's estimate.

For model choice, ask the agent which model it plans to route each shot to and the credit weight of each. invideo holds the full current roster — Veo, Kling, Seedance 2.0 for video; Recraft, Nano Banana, GPT-Image-2 for image — and the agent picks per shot, so cost varies by model. A quick "give me a sitrep: which models, how many clips, total credits, dollar equivalent at my plan rate" pulls the full pre-flight number into one message.

If the estimate is higher than your budget, tell the agent — it will cut clip counts, drop to a lighter model, or switch to 720p Seedance 2.0 renders (significantly faster and cheaper than 1080p, and visually sufficient for vertical formats) before regenerating the plan.

Watch some of these to see what works for you:

It also gives me an estimated generation breakdown, which is how many reference sheets, how many video clips, how many UI screens it's planning to generate, which is great because I know what I'm roughly going to spend now before I actually spend it.

— invideo's creative team, documenting the pre-flight estimate step