Enterprise Recommendation Engine for Broadcasters, Telecoms and Pay TV
For operators whose catalog, brand and procurement process rule out a self-serve API key. Your content trains the engine, your brand fronts it, and the terms go in a contract.
TasteRay Enterprise puts an Emotional AI recommendation engine inside your own service, for broadcasters, telecoms and pay TV operators. It covers custom AI trained on your content, white-label delivery, an SLA, dedicated support, analytics and custom limits. A DPA is available on request, and an engagement can start with a pilot.
A large operator asks different questions from a developer with a free key: who is accountable, what happens to viewer data, what a pilot proves. This page answers those. No training data is needed to start, and Enterprise can train on your content.
What Enterprise includes
- AI trained on your content
- The engine learns from what you hold, so picks and the reasons behind them reflect your catalog. We scope what training covers with you, because it depends on the content you can share.
- White-label
- The recommendation surface carries your brand inside your product. Your viewers deal with you.
- SLA
- Availability and response commitments are written into your contract. Self-serve tiers carry no SLA.
- Dedicated support
- Dedicated support in place of a shared inbox, working with your product and engineering teams.
- Analytics
- Analytics on how recommendations perform across your surfaces.
- Custom limits
- Limits set to your traffic, above the 200 requests a minute and 100,000 a month of the Pro tier.
Self-serve or Enterprise
Most teams can evaluate on a free key before any conversation with us. The contract starts to matter when you need your own content in the engine or a commitment you can hold someone to.
| Question | Self-serve API | Enterprise |
|---|---|---|
| Price | Free, $49 or $199 a month | Custom |
| Limits | Up to 100,000 requests a month, 200 a minute | Set to your traffic |
| Training data | None needed | None needed to start; can train on your content |
| Brand | Your interface, your copy around our text | White-label |
| Commitment | None | SLA in the contract |
| Support | Community, email or priority, by tier | Dedicated support |
| Analytics | Usage analytics on Pro | Analytics |
| Paperwork | Published terms and privacy policy | DPA on request, security review, contract |
How an engagement starts
Write to us
Tell us your catalog size, your territories and languages, and the surface you have in mind. An email is enough, and we reply with the questions that matter for your case.
Try it on your own titles
Before any paperwork, run ten of your own titles through the playground. It takes no key, and your editors can judge the reasons on content they know.
Scope a pilot
We agree one surface, one part of the catalog and one measure of success up front. Security review and the DPA happen here, so legal is not the last step.
Run the pilot
It runs on your product, in front of your viewers, for a group you choose, with a control group where you can have one.
Decide
If the measure moved, we write the contract: limits, SLA, support, and training on your content. If it did not, you have your answer.
What to check before you commit
- Set-top boxes and TV apps
- TasteRay ships a REST API, with no TV SDK. Your team or integrator builds the screens, and your backend calls the API.
- Response time
- Plan for a few seconds, with outliers. On a set-top box that means precomputed rows from your own cache, with live calls kept for when a viewer types or speaks.
- Household accounts
- The API stores no profiles, so you decide whether a request describes a person, a profile or a household. A pseudonymous identifier is enough, and often no identifier at all.
- Kids profiles
- The API privacy policy says the service is not intended for children under 13. Keep kids profiles out of the integration unless your contract says otherwise.
- Where it runs
- The API runs on Cloudflare Workers and sends request context to AI model providers. If your policy requires on-premises deployment, raise it in your first email.
Security and procurement
Our security page sets out the data flow, what is logged and for how long, and which providers see a request. A DPA is available on request.
SOC 2 Type II: in progress. ISO 27001: in progress.
Who you would work with
You would deal with people who have years of experience running product, technology and growth at scale. More on the team page.
- Mike Wolniak, CEO
- Built and sold the VML hub in CEEMEA.
- Mike Jaskolski, CTO
- Led product and technology for Nokaut and Morizon, both of which went public. Builds the TasteRay engine, trained on more than 500,000 long-form reviews.
- Marius Labudda, Growth
- Drove user growth at Sportano and G2A.com.
- Kordian Klecha, Product
- Has built products and services used by millions.
Frequently Asked Questions
Does TasteRay Enterprise need our data to start?
No. TasteRay needs no training data to start, because each request carries its own context, and a pilot can run that way on part of your catalog. Enterprise can also train on your content when you want the engine shaped by what you hold. We scope that with you first.
Can a broadcaster start with a pilot instead of a contract?
Yes, a pilot is a normal way to start with TasteRay Enterprise. We agree one surface, one part of the catalog and one measure of success before it begins, and the security review and any DPA happen during scoping. Write to us with your catalog size and territories to begin.
How much does TasteRay Enterprise cost?
TasteRay Enterprise pricing is custom and depends on your traffic, your limits and whether the engine trains on your content. For scale, the published self-serve tiers are free, $49 and $199 a month, the largest covering 100,000 requests. Ask us for a quote once you have scoped a surface.
Is TasteRay SOC 2 or ISO 27001 certified?
SOC 2 Type II is in progress and ISO 27001 is in progress at TasteRay. A DPA is available on request. Our security page sets out the data flow, the 90-day log retention and the role of AI model providers, so your team can review how requests are handled in the meantime.