Embedded AI analytics for your application
Embedding analytics is easy. Embedding analytics that cannot leak one customer's data into another customer's screen is the whole job.
A question becomes a decision
Can I ask this without leaving my application?
- 01
Connect
Your application + Nexatron SDK
- 02
Resolve meaning
Authenticated workspace and access scope
- 03
Review the answer
A governed answer in the workflow where the decision happens.
Illustrative workflow. Results depend on configured sources, definitions, and permissions.
Nexatron can be embedded in another product through the @nexatron/chat SDK, so an application's own users get natural-language answers over their own data. Tenant isolation and row-level rules are enforced where the query executes rather than in the embedding application, so the host product does not have to re-implement the security boundary.
What it does
- Isolation enforced at the query layerTenant scoping is applied in the database on the executed query. An embedding application cannot accidentally widen it, because it is not the thing enforcing it.
- A real client library, so you can compose it@nexatron/chat is a real client library with runnable examples, so the surface can be styled and composed rather than dropped in as an opaque frame.
- The same evidence, inside your UIEmbedded answers carry the same generated query, sources and confidence signal. Your users get the ability to check, which is what makes the feature retainable rather than a demo.
What it deliberately won’t do
Each of these is a case where returning something would have been easy and wrong, so the product is built not to.
- Scoped by the token you mintThe embed's reach is whatever the issuing token grants. This is deliberate: there is no ambient permission that survives a mistake in your integration code.
Test the claim, don’t take it
Ask it something you already know the answer to, then ask it something your data cannot answer. Those two questions tell you more than a feature list.