vidNT sits above the CDNs you already have and answers one question, continuously: for this viewer, on this network, right now — which of your CDNs should serve the next segment?
"Which CDN is best for Germany" is not a real question. The real one is "which CDN is best for this viewer, on this mobile network, in this city, during this event".
Decisions are scored per geography and per access network, because a fibre viewer and a mobile viewer in the same city routinely need different answers.
The engine re-ranks your CDNs on a short cycle from live signals — not from a table someone edited last quarter.
Every decision carries the reason it was made. You can replay why a viewer went where they went.
The CDN for a viewer is picked before playback starts, from what we already know about their geography and network. Then the brain stays on that session: it keeps reading what the viewer is actually getting, and moves them the moment their CDN starts to fail them. Most systems decide once, at the first segment, and never revisit — so a live event that goes wrong in the middle stays wrong for the rest of the stream.
Nothing. The change lands on a segment boundary, so playback continues without a stall, a re-buffer or a visible quality drop.
The player's own reported buffer level and bitrate, plus what the serving CDN says about itself — not a guess from a synthetic probe in a data centre.
Session affinity and no-switch windows stop the system from flapping, and SCTE-35 markers are respected so a switch never lands in the middle of an ad break.
The player already reports how it is doing. The CDN already answers back. Nothing here is exclusive to us — and nothing here is a decision. The decision is ours: for this viewer, on this network, right now, which CDN feeds the next segment.
Common Media Client Data. The player attaches what it is actually experiencing to each request, so the routing decision is made from the session rather than about it.
Common Media Server Data. The CDN answers back on its own responses, which is the only first-hand account of how the delivery side is actually holding up.
vidNT decides for every single session and still adds zero delivery risk: not one byte of your video passes through us. Your CDNs serve every viewer, exactly as they do today.
Not us. The playback path never passes through vidNT.
Into DNS and into Content Steering, so the answer is already in place when a player asks.
The last published decision stays in force. You lose new decisions, not the stream.
The product is the same either way. What changes is who operates it and where your viewers' data comes to rest.
We operate the platform. Nothing to install, nothing to patch. The fastest way to get a first service routing.
The platform runs inside your own cloud account. Your keys, your retention policy, your network boundary.
The classic path. Connect the CDN accounts you already have — your contracts, your rates, your relationships stay exactly as they are. vidNT stitches across them.
No CDN contracts yet? One agreement with vidNT, and delivery runs on sub-accounts we hold with leading CDNs — priced by our combined volume, stitched by the same AI brain. Start on one, add more as you grow.
Whether that is a multi-CDN vendor, weights in a DNS console, a list inside your player, or something your team wrote — you should not have to retype it.
Three things have to be true for steering to work at all. All three of them are already yours.
VIDNT works with the providers you are already contracted with. Which ones we can drive directly, and how deeply, is a live list — ask us for the current matrix rather than trusting a logo wall.
Anything that can emit CMCD and follow Content Steering: common web players, native players, and your own if it speaks the standards.
The Content Steering manifest is published through the DNS you already run, so there is no new authoritative service to migrate to.
Demand arrives faster than any single CDN was provisioned for. When one starts to strain, VIDNT moves viewers individually — including sessions already playing — before anyone notices.
Relevant capability: mid-stream switching.
Your dashboards are green and your viewers are complaining. That usually means the problem lives in one network, in one city — exactly the resolution a regional average erases.
Relevant capability: region × network routing.
You hold committed volumes across several CDNs and no practical way to hit them deliberately. VIDNT can weigh cost alongside quality — and shows you the trade before it takes it, never after.
Cost is an input to routing. It is never billed by us — see pricing.
A second CDN only helps if something can actually move traffic to it while an event is live. A contract is not a failover plan.
Relevant capability: out-of-path publishing.
The fastest honest answer is a comparison with your own stream, not a feature table.
We route between your CDNs, pass your DRM through untouched, and answer whatever player your viewers already use — no SDK and no plugin in your player.
CDNs we route between
DRM & security we pass through
Players we tune for
Any player that speaks HLS or DASH works. The names above are the ones we detect and tune for by default.
No. VIDNT decides which of your CDNs serves each viewer; the CDNs still do the delivering. If you run a single CDN today there is nothing for it to choose between — the first step is a second provider, not a new product.
Playback continues. Decisions are published into DNS and Content Steering ahead of the request, so the last published answer stays in force. You stop getting new decisions; you do not stop delivering. How that works.
Usually a configuration change rather than a code change — most modern players can emit CMCD and follow Content Steering once they are told to. Where a player cannot, a one-line embed turns CMCD on from outside it.
A weight moves a share of sessions and can see none of them. It cannot tell that one mobile network in one city is struggling, and it cannot move a viewer who has already started playing. Per-viewer routing · mid-stream switching.
That is what the deployment choice decides. Run it in your own cloud and raw session data never leaves your account, under your keys, on your retention policy. Both deployment models.
Ask us for the current matrix. We deliberately do not print a provider list on this page: naming one implies a relationship, and the depth of control genuinely differs between them.
CMCD, CMSD, Content Steering and SCTE-35 — what each one carries, and what we do with it.
DNS weights, a vendor, or building it yourself — side by side, including where each one is the right answer.
Paste a playback URL and read what your delivery looks like today. No account, no credentials.
Whoever you are contracted with today. We add to them; we do not replace them.
They emit CMCD — often with a one-line change, sometimes with none at all.
Scores every CDN per region and network, continuously, from live signals.
Writes those decisions into DNS and Content Steering, ahead of the request.
A decision per viewer, per chunk
Learns in-server from every session
Your CDNs serve the manifests
If vidNT stops, playback continues
No new single point of failure
CMCD · CMSD · Content Steering · SCTE-35
Works with players as they are
No proprietary plugin, ever
Full REST API — the console is just another client
OpenAPI docs · scoped API keys · webhooks on every event
MCP interface for AI agents (beta)
Signed URLs survive the mid-stream switch
AES-128 with authenticated key delivery
SSO (OIDC) · immutable audit logs
Fleet → region → network → session → chunk
Every routing decision accounted for
Privacy rules built into the pipeline
Leave your details and we open your access. A person comes back to you within one business day.
Open your assistant with a ready prompt — ask it anything about multi-CDN stitching, how vidNT fits your stack, or how to start.
The prompt asks your assistant to read vidnt.com/llms.txt and answer from it.