FFmpeg API Pricing Compared: Every Option in 2026

Video processing pricing is confusing by design

Every FFmpeg API and cloud service prices differently. Some charge by GB transferred. Some by compute seconds. Some by API calls. Some have egress fees buried in footnotes.

This makes comparison difficult. A service that looks cheap at 100 videos might be expensive at 10,000. A service with a low base price might double your bill with egress fees.

This analysis normalizes everything to a common metric: cost per video processed at various volume levels. For feature comparisons rather than just pricing, see the 2026 FFmpeg API comparison.

The pricing models

Model 1: Command-based (RenderIO)

You pay per API call. One command = one price unit. File size doesn't matter. Zero egress fees.

Cost per command (base plan):

• Starter: $0.06

• Growth: $0.029

• Pro: $0.0098

• Business: $0.00495

Model 2: GB-based (Rendi)

You pay per GB of data transferred (input + output combined), plus vCPU allocation.

• Free: $0/mo (50GB processing, 4 vCPUs, 1 min runtime)

• Pro: $25/mo (100GB+ processing, 4-256 vCPUs, up to unlimited runtime)

• Enterprise: Custom

Cost per GB on Pro (at 100GB): $0.25

For a typical video (50MB in, 30MB out = 80MB total):

• Cost per video on Pro: $0.02

For a large video (500MB in, 300MB out = 800MB total):

• Cost per video on Pro: $0.20

Model 3: Compute-time based (AWS Lambda)

You pay for compute seconds, storage, and data transfer separately.

• Lambda compute: $0.0000167/GB-s

• API Gateway: $3.50/million requests

• S3 storage: $0.023/GB/month

• S3 egress: $0.09/GB

Model 4: Container-time based (GCP Cloud Run)

You pay for CPU and memory per second, plus storage and egress.

• CPU: $0.000024/vCPU-second

• Memory: $0.0000025/GiB-second

• Cloud Storage: $0.020/GB/month

• Egress: $0.12/GB

Model 5: Fixed cost (Bare metal VPS)

You pay a flat monthly rate regardless of usage.

• Hetzner CPX31 (4 vCPU): ~$15/month

• DigitalOcean (4 vCPU): ~$48/month

Cost per video at different volumes

Assumptions for comparison

• Average input video: 50MB, 30 seconds, 1080p

• Average output: 30MB

• Average processing time: 10 seconds of compute

• Lambda: 1GB memory, 10s execution

• Cloud Run: 2 vCPU, 2GB RAM, 10s execution

100 videos/month

*Hetzner includes 20TB of traffic.

Winner at 100 videos/month: Lambda (cheapest raw cost) or RenderIO ($9/mo, zero ops). If you go the Lambda route, read about the limitations of serverless FFmpeg first.

1,000 videos/month

Winner at 1,000 videos/month: Lambda (cheapest raw cost) or Hetzner (no ops overhead for the price)

But Lambda requires 2-8 hours of setup and ongoing maintenance. Factor in engineering time at $150/hour and RenderIO's$29/month is cheaper than 12 minutes of engineering time per month.

10,000 videos/month

At 10,000 videos, you need 2 Hetzner VPS instances to handle the concurrent load. RenderIO's Pro plan covers 5,000 commands with overage at $0.03/cmd, or the Business plan covers up to 20,000 commands at$99/mo flat.

Winner at 10,000 videos/month: RenderIO Business ($99, no ops) or **Hetzner** ($30, significant ops)

50,000 videos/month

At 50,000 videos, self-hosted Lambda and Cloud Run become expensive due to egress fees. RenderIO's Business plan covers 20,000 commands at $99/mo with overage at$0.02/cmd for the remaining 30,000.

Winner at 50,000 videos/month: Hetzner ($120, significant ops) or **Lambda** ($276, moderate ops). At very high volume, contact RenderIO for custom pricing.

Hidden costs

Egress fees

The biggest hidden cost in video processing. Every time you download a processed video:

• Cloudflare R2 (RenderIO): $0

• AWS S3: $0.09/GB

• GCP Cloud Storage: $0.12/GB

• Azure Blob: $0.087/GB

At 10,000 videos with 30MB average output, egress costs:

• RenderIO: $0

• AWS: $27

• GCP: $36

Egress costs scale linearly. At 100,000 videos, AWS egress alone is $270/month.

Storage costs

Processed files need to live somewhere until downloaded:

• Cloudflare R2: $0.015/GB/month (included in RenderIO plans)

• AWS S3: $0.023/GB/month

• GCP Cloud Storage: $0.020/GB/month

Operational costs

Self-hosted solutions require engineering time:

This is the cost that makes the raw infrastructure comparison misleading. The FFmpeg as a service model exists precisely because most teams don't want to spend engineering time on video infrastructure.

Cold start costs

Lambda cold starts add 5-10 seconds to the first invocation after idle. If your video processing is bursty (common in batch workflows), cold starts affect user experience.

RenderIO's edge workers have minimal cold start because Cloudflare keeps workers warm across their network. The complete FFmpeg API guide covers how the request flow works in practice.

Break-even analysis

When does self-hosted become cheaper?

Comparing RenderIO plans to self-hosted (Lambda or bare metal):

Raw infrastructure only (ignoring engineering time):

• Lambda breaks even with RenderIO's Growth plan ($29/mo) at ~5,200 videos/month

• Hetzner is $15/mo, always cheaper on raw cost alone

Including engineering time ($150/hr, 10 hours/month):

• Lambda never breaks even below 100,000 videos/month

• Hetzner breaks even at ~80,000 videos/month

The engineering overhead makes managed APIs cheaper for most teams up to very high volume.

Recommendations by use case

Startup/indie developer (0-500 videos/month)

RenderIO Starter plan ($9/month)

500 commands for $9/mo. Zero egress fees. Zero ops burden means you focus on your product, not infrastructure.

Growing SaaS (151-1,000 videos/month)

RenderIO Growth plan ($29/month)

1,000 commands for $29/mo with overage at$0.05/cmd. No egress. No scaling concerns. The cost is a rounding error compared to your other infrastructure expenses.

Scale-up (1,001-5,000 videos/month)

RenderIO Pro plan ($49/month)

5,000 commands for $49/mo with overage at$0.03/cmd. Still cheaper than the engineering time of running your own infrastructure.

High-volume (5,001-20,000 videos/month)

RenderIO Business plan ($99/month)

20,000 commands for $99/mo with overage at$0.02/cmd. Predictable pricing at scale with 200 GB storage included.

Enterprise (20,000+ videos/month)

RenderIO Business + overage or custom plan

At this volume, the Business plan with $0.02/cmd overage remains competitive. For very high volume, talk to RenderIO about custom pricing. Or evaluate self-hosted if you have a dedicated infrastructure team already managing video workloads.

The real question

The cheapest option depends on what you value:

• Cheapest raw cost: Self-hosted on Hetzner ($15/month)

• Cheapest total cost (including engineering time): RenderIO (plans from $9/month)

• Most predictable cost: RenderIO (fixed per-command pricing, zero egress)

• Most flexible: Self-hosted (full control)

For most teams processing video, the managed API is cheaper once you factor in engineering time. Self-hosting makes sense if you already have a team managing similar infrastructure and your volume justifies the fixed costs.

Don't optimize for the cheapest server bill. Optimize for the cheapest total cost.

If you want to get hands-on, the REST API tutorial walks through making your first call. We also have guides for Python and Node.js. For a deeper look at self-hosted tradeoffs, see hosted FFmpeg vs self-hosted.

Plans start at $9/mo. See what the FFmpeg API includes or get your API key to start processing video today.