All-in COGS per minute: Spot GPU (1m render + 2m boot) + LLM + TTS/Whisper + R2 Storage + Tornado vs. Vitruvian API. See sources & verification at the bottom.
30m = 0.90 GB
Published floor is $0.06/GB (≈ $0.0018/min at 30 MB/min). The $0.0067 default is an assumption. Confirm your real rate in the Tornado dashboard.
Unverified: no public pricing page found. Treat $0.06/min as a placeholder until the real quote is in.
Cold start + teardown
1:1 active GPU encode
Amortizes boot time
-
-
-
-
Where every Naira goes per 1 minute of generated video before profit margin.
| 1. Tier Pricing & Allocation (Editable) | 2. Shared Infra Stack | 3. All-In Tornado | 4. All-In Vitruvian | 5. Tornado Advantage | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Tier Name | Price (₦) | Credits (Mins) | Rate/Min & Paystack | Spot + LLM + Stack | All-In Cost ($ / ₦) | Net Profit & Margin | All-In Cost ($ / ₦) | Net Profit & Margin | Saved w/ Tornado | |
Margin badges highlight profitability after deducting all costs:
Tornado: - • Vitruvian: -
Tornado: - • Vitruvian: -
Tornado: - • Vitruvian: -
Checked November 2026. "Verified" means a public source confirms the number. "Estimate" means it depends on our own usage and should be measured from real jobs.
| Item | Used here | Status | What we found |
|---|---|---|---|
| Tornado API | $0.0067/min ($0.222/GB) | Partly verified | Billed per GB, not per minute. Published rate is "as low as $0.06/GB" with a volume commitment (custom above 30 TB). New accounts get 100 GB free for 7 days. A billing floor can charge short clips or audio at the source's reference size, so billed volume may exceed the file size. At 30 MB/min the best case is about $0.0018/min. Check the real rate in the dashboard. tornadoapi.io/pricing |
| Vitruvian API | $0.06/min | Unverified | No public pricing page for a video API with this name could be found. The $0.06/min figure is a placeholder. Replace it with the vendor's quote. |
| Spot GPU | $0.55/hr | Verified | AWS g5.xlarge (1x A10G) in eu-west-2, the region our render workers use: spot about $0.47 to $0.58/hr, on-demand about $1.28/hr. Spot disruption is above 20%, so some jobs will retry or fall back to on-demand. DoiT price tracker |
| Boot and render time | 2m boot + 1m render | Estimate | Measure from real render logs. Boot cost is paid per job, so longer videos or a warm worker cut it sharply. |
| LLM tokens | $0.008/min | Estimate | Depends on the model and prompt size. Compute from actual token counts per job times the provider's per-token price. |
| TTS + Whisper | $0.003/min | Estimate | The repo ships local faster-whisper models and a local TTS environment, so if both run on the same GPU their cost is already inside the Spot line. If so, this line double counts and can be lowered or unticked. Hosted TTS would cost more. |
| R2 storage | $0.002/min | Estimate | R2 charges for stored GB-months and operations, with no egress fee. A 30 MB minute held for a month is a small fraction of a cent, so the line is mostly a buffer for retention time and queue cost. |
| Paystack | 1.5% + ₦100, cap ₦2,000 | Verified | ₦100 is waived under ₦2,500, fees are capped at ₦2,000 per transaction. paystack.com/pricing |
| FX rate | ₦1,600/$ | Assumption | Set to the current market rate. Every Naira figure scales with it, so test 1,400 to 1,700. |