Everything owners ask before they switch

The platform on your own server, so you can test the engine against your real streams, your real hardware and your real customer load. The trial includes everything except AI service and GPU processing; those two are part of the Premium plan. Nothing else is held back.

Ubuntu 20.04, 22.04, 24.04 or 26.04 for the panel and streaming servers — 26.04 is recommended. GPU servers need 22.04 or later, and every future Ubuntu release is supported as it arrives.

One server is enough to begin. You add streaming servers and GPU servers as the audience grows, and there is no licence limit on how many of either you run. What you actually need depends on your line-up and how much of it you transcode — tell us what you're running today and we'll go through it rather than publish a number that won't match your situation.

No. The platform installs and configures each server itself, including NVIDIA drivers and DKMS on GPU machines. Adding a server is a decision, not a weekend.

Up to 1 million viewers. The database is built to keep answering through millions of requests, which matters most at the moment your whole customer base connects at once — a match kicking off, not three in the morning.

Every competing panel puts a web server in front and a script interpreter behind it, both sitting in the path of every stream request doing work unrelated to video. We wrote our own HTTP server for this one job instead. In practice it means lower load per stream — around 30–40 MB of RAM for a running stream — so one machine carries more, and the ceiling everyone else shares stops being yours.

A channel that runs only while somebody is watching it. It starts when the first viewer arrives and stops after the last one leaves, so a large line-up costs a fraction of running everything around the clock. The usual objection is the wake-up delay; ours is about 800 milliseconds, which is short enough that the viewer never learns the channel wasn't already running.

Yes — zapping is immediate. It's also the thing customers judge you on within five minutes of signing up, and the hardest thing to fix later when the core belongs to somebody else.

No. The old Xtream API is supported, so the apps your customers already use keep working exactly as they do now. Nothing to rebuild, nothing to resubmit to an app store, nothing for your customers to reinstall.

Instead of reading out a server address, a username and a password to somebody typing them on a remote control, the panel issues one short active code. The customer enters that code and the device configures itself against your panel, with the right packages, device limits and expiry already set by you.

Not to get set up. The account details behind the code are configured by you in the panel, so there is nothing to dictate over the phone and nothing for the customer to mistype.

Yes, the MAG device system is part of the platform, alongside the widely used streaming apps and players your customers already own.

Two things you control on the account: the number of devices allowed to watch at the same time, and the record of every activation. Between them, an account being used well beyond a household is visible rather than invisible. There is also protection against account sharing and unusual viewing patterns.

Yes, within the Premium plan. SolusTV's built-in AI comes with a free API and free token usage, and needs no key and no account. There is no per-minute or per-token charge, because the models run on your own hardware rather than on somebody else's service.

Yes. Paste your key in and the work is billed to your provider account instead. The key is stored encrypted on your own server and you can remove it at any time. It's optional — most owners never need it.

Yes. The models download and install themselves onto the GPU server you choose, along with the whole environment — Torch, CUDA and Whisper — from one button. Nothing leaves your building.

That's your choice. Five speech model sizes are available, from Tiny to Large-v3; larger models are more accurate but heavier. A practical approach is a smaller one to work through a back catalogue and a larger one for the titles you promote.

A dedicated 200-language translation model handles it, with Gemma AI available as an alternative. You don't tell it what language the film is in — the source is detected automatically and you only choose the target.

Not between jobs. Translation models are loaded by the job that needs them and freed as soon as it finishes, so there is no permanent AI service holding memory on your card.

NVIDIA, and only NVIDIA. That's deliberate rather than a gap — supporting one vendor properly means we encode on CUDA, tune against real NVIDIA behaviour and can tell you what a given card will carry, instead of shipping a lowest-common-denominator path across three vendors.

No. Drivers and DKMS are installed and configured automatically when you add a GPU server, and rebuilt through kernel updates. Any driver version is supported, so an older card you already own stays usable and a new card doesn't have to wait for us.

Unlimited, with no per-card surcharge. Adding encoding capacity is a hardware decision and nothing else.

Yes — a single input is encoded into multiple resolutions at the same time, so the viewer on mobile data and the viewer on fibre are both served properly from one source. You can also put your own logo on a channel, or remove one that came in on a source.

Resellers buy credits and spend them creating or renewing customers. Nobody can run up a debt with you, because the balance has to exist before the account does. Every movement is on the record with a date against it, so a disputed balance is a question with an answer.

No. Reseller packages define exactly what each partner may sell, and credits can only be spent within that. The limits are set by you, not negotiated by them.

Yes. A reseller can build sub-resellers underneath them, so your network grows without every new seller becoming yours to manage. Every level still operates inside the package you defined at the top.

A reseller can present the service under their own name and their own domain, managed from their portal. Their customers see their brand; you keep the platform, the limits and the record of everything underneath.

Yes — a separate reseller portal, distinct from the owner's panel, with its own currency and presentation settings, plus live chat and tickets built in so a waiting partner reaches you where the work is.

Yes. A rule can act on what an address is — a datacentre or hosting address, a VPN exit node, an open proxy, a Tor exit node or a known abuser — as well as where it comes from, including the AS number and the network operator behind it.

That's why there are signals for what not to block: corporate and privacy relays, verified crawlers, mobile networks and satellite networks. "Possible VPN" is offered separately from a confirmed VPN exit node, so you can be aggressive where you're certain and cautious where you aren't.

Several things, deliberately. An address holding a valid line is never blocked automatically. A refused connection is turned away before a session exists, so it never uses one of the user's allowed connections. If a country can't be determined the rule is skipped rather than enforced. And you can test a real address against your rules before anything depends on the answer.

No, it's optional and off by default. When you do turn it on it starts in monitor mode, blocking nothing and simply recording what it would have done — including how much of that would have been your own customers. You enforce only when the evidence from your own server says you should.

Yes. Staff accounts have precise permissions, so somebody handling support doesn't reach billing, servers or content. Every action appears in a full audit, and the owner's account can require two-step sign-in.

Yes, and migration is free. Channels, films, customers and packages are imported from your existing setup, and you can run both in parallel during the transition rather than switching everything on a single evening.

No. Run SolusTV V2 alongside your current platform during the three-day trial and migrate properly afterwards if it does what you need.

They shouldn't have to do anything. The old Xtream API is supported, so existing apps keep working, and new devices can be set up with a single active code instead of a settings screen.

Yes. There's a documented API for creating, renewing and managing customers, suited to connecting a billing system or your website so a sale becomes an account without anyone retyping it.

€100 for your first month, then €200 a month for the full platform — unlimited streaming and GPU servers, AI service with unlimited tokens, the active code and MAG device systems, firewall, resellers, TV archive, Cloudflare setup, daily backups, free migration and 24/7 support. Before that, a free three-day trial.

No per-server fee, no per-card fee, no metered AI, no charge for migration and no charge for support. The things other panels bill you for afterwards are in the licence.

If it’s doing what you need, your first month is €100 and it renews at €200 a month, with AI and GPU switched on and everything carrying straight over. If it isn't, tell us why — that's the more useful outcome for us.

Yes — live chat and tickets, 24 hours a day, at no extra cost. Messages are translated automatically in both directions, so you write in your own language and are answered in it, and the full conversation history is kept.

On hardware you own or rent. Your channels, library, customer list, resellers, firewall rules and logs all stay on your own servers. Nothing about your business sits with us.