Skip to the content
Novixo

Literature Acquisition

One controlled route from your AI agent to the literature

Acquisition and evidence infrastructure for pharmacovigilance literature. Your own AI agent, application or existing system acquires literature from many sources through one interface, and the origin and evidence of every record are kept.

Available for pilot projects

What it is

Infrastructure, not another PV application

It does not replace your pharmacovigilance platform and it does not decide anything. It answers one question well: what has been published about this product, where did each record come from, and which sources could not answer today.

  • Works with the AI assistant or agent you already use
  • Hands results to the systems you already have
  • Keeps what was seen, where and when, for every record
  1. Your AI agent, application or existing system

    Connects through the Model Context Protocol or a REST API.

  2. Access layer

    Per-customer API keys, scopes, plans and quotas. The customer is bound before any data is read.

  3. Acquisition engine

    Plans each search per source, calls the source by its own route, parses the records, and records a reason for every source that did not return everything.

  4. Identity and de-duplication

    One record per article, with every source it was seen in kept. Provenance and evidence are hash-checked.

  5. Your database

    Results are delivered into your own systems through a versioned contract, with a receipt read back.

How it behaves

Built for the question an auditor will ask

“How do you know nothing was missed?” starts with knowing exactly what was searched, and what could not be.

Many sources, each by its own route

Publisher and index APIs, OAI-PMH, RSS and Atom feeds, journal sites. Each source is reached the way it is meant to be reached.

An answer you can rely on

Every source ends in exactly one outcome with a reason. A source that failed is never reported as “no results”.

Authorised access only

No paywall, login or anti-bot barrier is ever circumvented. A source behind a licence you do not hold is reported as such.

Sources are watched

A monitor probes each source, keeps its health, notices when a source changes or goes down, opens an incident and checks again after a fix.

Customers kept apart

A customer-scoped data layer and row-level security on every table. Another customer’s identifier answers as if it did not exist.

Your AI is the interface

The infrastructure runs with no model of its own. Your agent asks; it acquires, keeps the evidence and delivers.

Building an agent that reads the literature?

Tell us which products and sources matter to you. We will show you what the infrastructure returns for them, including what it reports when a source cannot answer.