Estimating cost for a large or long-form project
Real numbers for budgeting a project with a lot of footage — what import actually costs, ten specific paid actions with real rates, and a worked example table.
If you're weighing Eddie against a budget — a documentary, a long interview series, months of footage — "it depends" is a true answer but not a useful one on its own. This walks through what actually drives cost, with real numbers, so you can build your own estimate instead of guessing.
The two things that cost credits
Importing footage costs 250 credits per hour of source, billed in 15-minute blocks (15-minute minimum — the smallest import is 63 credits). This is the one you can size up front, because it scales directly with how much footage you have.
Editing is mostly free. Cuts, trims, captions, titles, b-roll placement, color, motion — most of what you'd prompt Eddie to do costs nothing beyond the import.
A specific set of AI actions do cost credits, each priced by the actual unit of work — minutes of audio or video generated, characters of script, hours of footage analyzed, or clips produced — not a flat fee per action. Ten examples, at current rates:
| Action | What it costs | Example |
|---|---|---|
| Generate AI music | 200 credits/minute generated | an 80-second track ≈ 267 credits |
| Generate an AI voice-over | ~178 credits per 1,000 characters (~150 words) | a 300-word script (~2,000 characters, the max per call) ≈ 356 credits |
| Generate AI b-roll (video) | 156 credits per 8-second clip (standard tier) | a 3-clip sequence ≈ 468 credits |
| Cut b-roll subclips from your own footage | 600 credits/hour analyzed, 10 credits/clip minimum | an hour cut into 12 clips ≈ 600 credits |
| Stabilize shaky footage | 40 credits/minute of source, per clip stabilized | a 90-second handheld shot ≈ 60 credits |
| Re-transcribe a source | 250 credits/hour, 15-minute minimum | a 10-minute re-transcribe ≈ 63 credits |
| Detect a music track's beat grid | 2 credits/minute of audio | a 3-minute track ≈ 6 credits |
| Get Eddie's editorial feedback on a cut | ~4.3 credits/minute + ~6.5 fixed | a 10-minute cut ≈ 50 credits |
| Resync all multicam angles automatically | ~2–3 credits/minute of ungrouped footage + ~6 fixed | an hour of ungrouped footage ≈ 150 credits |
| Export a tiered selects timeline | ~15 credits/hour of footage in scope | 10 hours in scope ≈ 150 credits |
A large first-assembly build across a lot of footage still costs more than a small tweak even when no single action above is involved — that part is priced on the work the run actually does, not listed here as a rate. This side is harder to size in advance than import, because it depends on what you ask for, not just how much footage you have.
Worth being specific about b-roll, since "add b-roll" sounds like one thing and is really several: finding it (search), reading what's already in your footage (analysis), and placing it on the timeline are all free — analysis runs automatically at import, which already paid for it. The only paid step is creating footage that doesn't exist yet: generating new AI b-roll, or cutting fresh subclips out of raw footage with AI guidance (both listed above).
Full detail on both, including exports: How credits and billing work.
Worked example: importing your footage
Here's the import cost at a few sizes, at the published rate of 250 credits/hour:
| Footage | Import credits | Roughly, at list pricing |
|---|---|---|
| 1 hour | 250 | free (inside your first 1,000) |
| 10 hours | 2,500 | ~$23 |
| 50 hours | 12,500 | ~$100 |
| 100 hours | 25,000 | ~$175 |
| 800–1,000 hours | 200,000–250,000 | talk to sales |
That's import only — before a single prompt, edit or export. The "roughly" figures use the per-credit rate of whichever pack fits closest ($10/1,000, $45/5,000, $120/15,000, or $350/50,000 — see Pricing); credits never expire, so buying a bigger pack than one project needs isn't wasted — the rest carries over.
What imports don't tell you about the rest
Import is the predictable half. The other half — the AI editing work itself — is priced on what each run actually does, and for a project this size that's genuinely hard to predict from footage volume alone. A first assembly across months of raw footage costs more than fixing a caption; how much more depends on the edit, not a formula. The most honest way to find out is to run the numbers on your own footage rather than trust an estimate built on averages that may not look like your project.
Exports
Sending a finished timeline to your NLE costs 100 credits per export, flat, whatever format — Premiere, Resolve, FCPXML, OTIO, EDL. A multi-episode series exported once per episode still adds up to a few hundred credits at most — a rounding error next to the import cost above. Re-exporting after notes costs the same 100 each time.
Try it on real footage before you scale up
New accounts start with 1,000 free credits — enough to import about 4 hours and take a real cut through to export. For a project this size, that's the useful move: run one episode, chapter, or day of footage all the way through — import, prompt, export — and read your own actual burn in the sidebar. That's a real number from your own footage and your own editing style, not an average from someone else's project.
Team projects
If a team is cutting this together, credits are drawn from one shared workspace pool, not billed per person. See What everyone in a workspace can see (and spend) before you add editors to a workspace with client footage in it.
Related: How credits and billing work, Import your footage, Contact support.