A live-video bill has separate layers: Callaba software, compute, viewer delivery, and storage. Model them together here, then choose AWS Marketplace or a self-hosted Linux deployment with each charge shown separately.
See self-hosted optionsMarketplace usage covers the Callaba software dimension. AWS compute, disks, network traffic, CDN delivery, and storage remain separate AWS charges.
Use your own Linux compute, network, storage, and delivery contracts. The Callaba license is calculated separately from those infrastructure costs.
Choose the nearest operating pattern. The calculator fills in runtime, audience, restream, and archive values so you can replace them with measured numbers instead of starting from an empty form.
Weekly runtime is modest, but several social outputs and a replay archive still create delivery, restream, and storage work.
Peak audience and watch time make CDN delivery much larger than the ingest line. Recording adds a smaller but persistent archive bill.
Long viewing sessions, sharp audience peaks, several destinations, and replay retention all matter on game day.
A 24/7 channel makes instance-hours predictable. Audience delivery stays variable, and an unnecessary archive can become the quiet long-term cost.
A smaller private audience reduces CDN pressure while network ownership, access policy, and data location become more important.
Runtime, playback, archive, and delivery need to fit the usage model of the product that calls the API.
The estimate separates Callaba software or license charges from infrastructure, CDN, and storage. Open the advanced view when you need to change quality or providers.
Enter the hours, average concurrent viewers, outputs, and retention. The result keeps Callaba charges separate from AWS or self-hosted infrastructure, CDN delivery, and storage.
Use this link to share the same workload assumptions. AWS instance runtime and network egress are calculated separately.
EU/NA standard delivery
The modelled subtotal includes the calculated Callaba units and selected CDN and storage assumptions. It excludes your Linux compute, disks, network, redundancy, operations, support, tax, and production capacity validation.
A baseline software assumption for the selected running hours, not a Callaba capacity recommendation. Validate the real codec, delivery profile, and infrastructure before deployment.
Marketplace software plus the selected provider assumptions for delivery and storage. Excludes compute, disks, network egress, requests, tax, and production capacity validation.
A modelled subtotal for the selected modules plus provider assumptions. Your server, disks, network, tax, and production capacity validation remain outside this number.
Compare the estimate with the routing, monitoring, recording, and recovery work your workflow needs. The customer evidence below describes the complete operating path.
The verified customer reported lower latency, greater reach, faster time to market, higher engagement, and lower infrastructure and operational costs for one live-events workflow.
Corporate events and webinars with instant recording, gallery views, monetization, distribution, low-latency delivery, CDN support, and collaboration across production teams.
The customer reported a 33% latency reduction after consolidating the live-video workflow.
Distribution and playback scale when the same live source can feed browser playback and multiple public destinations together.
The reported saving matters only with the workload, region, and operating changes that produced it; it is evidence, not a universal discount.
Audience hours drive delivery, instance-hours drive AWS software usage, and retention drives storage. Restream outputs and simultaneous jobs affect both capacity and self-hosted units.
Average concurrent viewers multiplied by live hours determines the modeled viewer-hours. Bitrate turns that into the transfer billed by the CDN.
A 24/7 channel keeps the Marketplace dimension active all month. Event deployments can stop the instance when the work is finished.
The estimator calculates one recording copy. Replication, request volume, retrieval, and a second recovery copy must be added separately.
These rates are planning inputs, not a quote. Region, traffic mix, requests, retrieval, taxes, commitments, and negotiated terms can change the invoice.
The calculator uses the Europe and North America standard-network rate. Other regions have higher public rates.
bunny.net CDN pricingA single public rate for high-volume 24/7 delivery, with a smaller point-of-presence footprint than the standard network.
bunny.net volume pricingA flat-rate distribution plan with bundled CDN, security, DNS, logging, storage credit, and a published 50 TB transfer allowance.
AWS CloudFront pricingA storage-only planning benchmark. Requests, retrieval, transfer, replication, and region can add separate charges.
AWS S3 pricingAlways-hot object storage with published free-egress conditions. The calculator models storage capacity only.
Backblaze B2 pricingUnits cover product capacity; your Linux host, network, CDN, and storage are separate. The calculator derives units for the selected streams, restream outputs, and recordings, then applies the published volume tiers to each module just as the subscription code does.
The selected streams, restream outputs, and recording jobs determine this calculator's unit subtotal. VOD viewers, pay-per-view viewers, and conference rooms add their own unit groups when used.
Bring a real monthly or hourly infrastructure quote. CPU, GPU, RAM, disks, traffic allowance, public addresses, backups, redundancy, and support are not included in the Callaba license.
A self-hosted origin does not make internet delivery free. Add the CDN or network-egress rate that applies to the audience regions and cache behavior.
Price the first copy, retention, backups, retrieval, requests, and transfer separately. Local disks and S3-compatible object storage fail in different ways and produce different bills.
Confirm which charges belong to Callaba, which belong to the infrastructure provider, and which depend on actual audience traffic.
Enter the weekly live hours, average concurrent audience, social outputs, and replay retention. The self-hosted side also counts the stream, restream, and recording units.
Use average concurrency across the live hours for the bill, then test peak concurrency separately for capacity. A short peak should not be multiplied across the entire event.
Model average audience for transfer, peak audience for acceptance testing, every simultaneous output, and the replay retention required by the rights agreement.
It fits when you already operate suitable Linux infrastructure or need private networking, a specific region, data control, or integrations that justify owning the runtime.
Yes. Each output adds a self-hosted restream unit and creates outbound traffic. The destination may not charge ingress, but your cloud or network provider may charge egress.
Viewer-hours expose the audience assumption; gigabytes expose the provider bill. The calculator shows both so an unrealistic audience number cannot hide inside a transfer total.
Yes. Keep the stream owners, outputs, storage, monitoring, and recovery tests documented so the deployment boundary can change without redesigning the operating workflow.
Launch in AWS when the Marketplace path fits. Continue with self-hosted planning when the server and network must remain under your control.
Launch the AMI in your AWS account, then verify the Marketplace software line and the AWS infrastructure lines separately.
Use the unit subtotal with an actual quote for your Linux compute, network, storage, redundancy, and support coverage.
Send the real channel count, audience, runtime, restream outputs, recording, retention, region, and recovery requirements for review.
