Success contract
One workflow, named inputs and outputs, observable pass/fail criteria, resource limits and an agreed stop condition.
+
+ MXL PRODUCTION PILOT / BOUNDED · MEASURED · REPEATABLE
+Turn an MXL question into a working production proof.
+A focused engineering engagement for broadcasters, vendors and integrators. We take one real workflow from contribution to control and delivery, measure it, document it and finish with a clear recommendation: go, adapt or stop.
+ +This is not a promise to replace an entire production platform. It is a disciplined way to decide where MXL creates practical value in your system.
+
+
+
+ 8native MXL contribution sources demonstrated in one production canvas
+ 2160p30CPU-first Program delivered to YouTube over RTMPS
+ 43.9–53.4%lower measured reader CPU than the compared NDI path in a bounded same-host test
+ <80 MiBfor the Core cut/take path in the measured reference configuration
+
+
+
+ 01 / THE OUTCOME
A decision package, not an endless prototype.
Every pilot starts with the decision your team needs to make and ends with enough evidence to make it. The working demo matters; the repeatable evidence and honest boundary matter more.
+
+ 01
+ 02
+ 03
+ 04
+ 05
+ 06
+
+ Working vertical slice
A real signal path using the minimum useful set of contribution, MXL processing, operator control and delivery functions.
Operator surface
A role-appropriate control or health view so the proof looks and behaves like broadcast equipment—not a collection of terminal windows.
Evidence package
Measured cadence, drops, CPU, memory and recovery behavior, plus screenshots, reproducible steps and known limitations.
Go / adapt / stop recommendation
A concise architecture recommendation showing what belongs in MXL, what stays at the edge and what should not be built yet.
Handoff
Public-safe documentation and a private operator handoff that another engineer can follow without depending on the original lab.
+
+
+ Compress at the edge.
+
+ 02 / STARTING PROFILE
Compress at the edge.
Share in the core.
The recommended first deployment is a single production host with enough CPU and RAM for the validated workload. SRT, NDI, SDI or ZoomISO can terminate at adapters; MXL carries timestamped media between in-host functions; a purpose-built edge encodes the finished output.
+
+ CONTRIBUTION
+
+ PRODUCTION
+
+ DELIVERY
+
+ ZoomISO · SRT · NDI
Keep transport-specific behavior at an adapter boundary. Prove only the sources the pilot actually needs.
MXL core + skin
Discoverable flows, cuts, layouts, audio, multiview and health on a measured CPU/RAM envelope.
YouTube · SRT · confidence
Encode once at the boundary and monitor the destination independently from the production core.
+
+ 03 / FIT CHECK
+Useful now—and explicit about what is not ready.
+
+
+
+
+
+
+ Good pilot fit
+-
+
- Build a credible MXL production demo for customers or internal stakeholders. +
- Integrate one vendor function or transport into an MXL workflow. +
- Compare a bounded MXL path with an existing NDI or SRT approach. +
- Measure a deployment before funding a larger product effort. +
Not the current promise
+-
+
- A feature-for-feature replacement for vMix, OBS, AMPP or a production switcher. +
- An unattended 24/7 appliance with a full broadcast SLA. +
- A universal claim that MXL always outperforms every NDI configuration. +
- A finished TAMS editing, trimming and replay product. +
+
+
+
+
+ 04 / WHY THIS LAB
The failures are part of the deliverable.
The reference work has already exposed cadence errors, stale flows, audio-boundary problems, cloud sizing limits and operator-recovery gaps. The pilot converts that experience into acceptance gates instead of rediscovering it on your budget.
+
+ BOUND THE CLAIM
+ REHEARSE THE INSTALL
+ OPERATE, DON'T JUST RENDER
+ KEEP THE GATE
+
+ Measure the exact path
CPU and reliability statements are tied to disclosed formats, readers, host classes and test conditions.
Fresh-host workflows
Deployment is treated as part of the product. A result that only works on the original lab is not accepted.
Health and recovery
The proof includes visible state, safe controls and recovery behavior—not only a successful frame grab.
No automatic expansion
The next feature or topology is funded only after the current pilot earns a reason to continue.
+
+ START WITH THE DECISION
+What would MXL have to prove for your team?
+Bring one workflow and one decision. We will define the smallest credible proof, the evidence it must produce and the condition that keeps the project from becoming open-ended.
+ +Scope, schedule, fee, required licenses and cloud costs are agreed before work begins. MXL is presented as complementary to NDI, SRT, ST 2110 and IPMX—not as a universal replacement for them.
+