What is Intelligence as a Service (IntaaS)?

IntaaS makes the intelligence behind each vGenix Ai module, such as validity checks, coding suggestions, literature screening, signal detection and E2B (R3) validation, available as REST APIs and as vGenix MCP servers, so you can use it inside the systems and AI agents you already run.

What is an MCP server?

The Model Context Protocol (MCP) is an open standard for connecting AI applications to tools and data. A vGenix MCP server lets an MCP-compatible assistant or agent call governed vGenix tools, read approved resources such as your SOP rules, and use approved prompts.

Should we use the API or the MCP servers?

Use the REST API when one of your systems, such as an intake portal or workflow tool, needs to call a capability. Use the MCP servers when an AI assistant or agent should use the capability. Both return the same result and evidence.

Can an AI agent act on its own through the MCP servers?

Only within the scope set for it in the Pharma Control Plane. Human gates your SOP requires still apply, and uncertain results are returned with a routing signal to send them to a person.

How is it different from the vGenix Ai platform?

The platform includes the five modules, workflow and workspace. IntaaS provides the intelligence only, and your team keeps working in your own systems.

Can we start with IntaaS and move to the platform later?

Yes. Both run on the same vGenixOS engine, so you can start with the service and move to the full platform, or use both.

Is IntaaS validated and audited?

Services are versioned and released under change control with validation evidence, every call is recorded, and every result returns its sources, rule and trace ID. You validate your own integration, supported by our evidence.

How is IntaaS priced?

[Confirm the IntaaS commercial model.]

vGenixOS / Intelligence as a Service

Intelligence as a Service · IntaaS

Intelligence
as
APIs & MCP servers

Take the intelligence behind every vGenix Ai module, from case validity and coding to signal evidence and E2B (R3) checks, and use it in your own systems through REST APIs, or in your own AI agents through vGenix MCP servers. Every call is governed by vGenixOS and every answer comes with its evidence.

Request sandbox access See the catalogue

Your systemssafety database, workflow, portals

→

REST APIJSON over HTTPS

Your AI agentsassistants, copilots, agent frameworks

→

MCP serverstools, resources, prompts

vGenixOS intelligence, governed

Same engine, rules and evidence as the vGenix Ai platform

Control planeNeurosymbolic checksHuman gatesAudit trail

The idea

Use the intelligence without changing your platform

Most safety organisations already have a safety database, workflow tools and, increasingly, their own AI assistants. Replacing them is slow and risky. IntaaS lets you add vGenix Ai intelligence exactly where it helps, one capability at a time.

From the platform

The same intelligence, unbundled

Each vGenix Ai module is built from capabilities: check validity, find duplicates, suggest coding, screen literature, detect signals, validate E2B (R3). IntaaS exposes those capabilities individually, running on the same vGenixOS engine as the platform.

Two ways in

APIs for systems, MCP for agents

Your systems call REST APIs. Your AI assistants and agents connect to vGenix MCP servers through the open Model Context Protocol, so they can use governed safety tools instead of guessing.

How Intelligence as a Service works: two ways in, one governed engine, and answers that carry their own evidence.

The catalogue

19 capabilities across five modules

Every capability is available as a REST endpoint and as an MCP tool. Each one draws on the four vGenixOS intelligences, Pharmacovigilance, Document, Regulatory and Compliance, and each one knows when to hand over to a person.

Capability	REST endpoint	MCP tool	Intelligence used	When uncertain or SOP-required
Classify a document	`POST /case/classify`	`classify_document`	Document	Automatic, logged
Check validity (four minimum criteria)	`POST /case/validity`	`check_validity`	DocumentPharmacovigilance	Routes to a person
Find possible duplicates	`POST /case/duplicates`	`find_duplicates`	Pharmacovigilance	Routes to a person
Extract case data with source links	`POST /case/extract`	`extract_case_data`	Document	Automatic, logged
Suggest MedDRA and product coding	`POST /case/code`	`suggest_coding`	Pharmacovigilance	Routes to a person
Check a narrative against the case	`POST /case/narrative-check`	`check_narrative`	PharmacovigilanceCompliance	Automatic, logged

Capability	REST endpoint	MCP tool	Intelligence used	When uncertain or SOP-required
Screen search results for relevance	`POST /literature/screen`	`screen_records`	Pharmacovigilance	Routes to a person
Identify cases in an article	`POST /literature/identify-cases`	`identify_cases`	DocumentPharmacovigilance	Routes to a person
Check EMA literature monitoring scope	`POST /literature/mlm-scope`	`check_mlm_scope`	Regulatory	Automatic, logged
Extract case data from full text	`POST /literature/extract`	`extract_article_case`	Document	Automatic, logged

Capability	REST endpoint	MCP tool	Intelligence used	When uncertain or SOP-required
Run disproportionality and trend detection	`POST /signals/detect`	`detect_signals`	Pharmacovigilance	Automatic, logged
Cluster related cases and terms	`POST /signals/cluster`	`cluster_cases`	Pharmacovigilance	Automatic, logged
Build a signal evidence pack	`POST /signals/evidence-pack`	`build_evidence_pack`	PharmacovigilanceDocument	Routes to a person

Capability	REST endpoint	MCP tool	Intelligence used	When uncertain or SOP-required
Reconcile listings with the source data	`POST /reports/reconcile`	`reconcile_listings`	Compliance	Routes to a person
Draft a section with source links	`POST /reports/draft-section`	`draft_section`	PharmacovigilanceRegulatory	Routes to a person
Verify statements against data	`POST /reports/verify`	`verify_statements`	Compliance	Automatic, logged

Capability	REST endpoint	MCP tool	Intelligence used	When uncertain or SOP-required
Determine reportability by market	`POST /submission/reportability`	`determine_reportability`	Regulatory	Automatic, logged
Calculate due dates from day 0	`POST /submission/due-dates`	`calculate_due_dates`	Regulatory	Automatic, logged
Validate an E2B (R3) message	`POST /submission/validate-e2b`	`validate_e2b`	RegulatoryCompliance	Automatic, logged

Endpoint and tool names are illustrative. The final catalogue, versions and availability are confirmed per agreement. [Confirm before publishing.]

Anatomy of a call

Every answer carries its own evidence

Whether your system calls the API or your AI agent calls an MCP tool, the answer looks the same: a result you can use and the evidence to check it. That is what makes IntaaS usable in a validated process.

ResultThe answer, structured: valid or not, the suggested term, the due date.

EvidenceThe source text, page and line behind every value.

RuleThe SOP or regulatory rule that was applied, with its version.

ConfidenceHow sure the engine is, against thresholds you set.

RoutingAccept, or send to a person when uncertain or required by your SOP.

TraceModel version and a trace ID that links to the vGenixOS audit trail.

RequestIllustrative

# Illustrative interface
POST /v1/case/validity
Authorization: Bearer <token>
Content-Type: application/json

{
  "document_id": "doc_017",
  "rule_set": "SOP-PV-012",
  "market": "EU"
}

Response200 OK

{
  "result": { "valid": true, "criteria_met": 4 },
  "evidence": [
    { "criterion": "identifiable_patient",
      "source": "page 1, line 6",
      "text": "67-year-old woman" },
    …
  ],
  "rule": "SOP-PV-012 v4 · GVP Module VI, VI.B.2",
  "confidence": "high",
  "routing": "accept",  // or "human_review"
  "model_version": "case-intel 3.2.1",
  "trace_id": "tr_8f2c41"
}

Connect your agentIllustrative

# Illustrative MCP client configuration
{
  "mcpServers": {
    "vgenix-case": {
      "url": "https://mcp.[tenant].vgenix.ai/case",
      "auth": "oauth"
    }
  }
}

Tool call and resultGoverned

# The agent calls a governed vGenix tool
tool: check_validity
arguments: { "document_id": "doc_017" }

# The same checkable answer comes back
result:   valid · 4 of 4 minimum criteria
evidence: source text for each criterion
rule:     SOP-PV-012 v4 · GVP Module VI, VI.B.2
routing:  accept   # uncertain → human_review
trace_id: tr_8f2c41

vGenix MCP servers

Give your AI agents safety expertise, with guardrails

The Model Context Protocol (MCP) is an open standard for connecting AI applications to tools and data. A vGenix MCP server gives any MCP-compatible assistant or agent three things, all governed by the Pharma Control Plane.

Tools

Actions the agent can call, each with its own scope and human gate.

check_validity
suggest_coding
calculate_due_dates
validate_e2b

Resources

Read-only context the agent can rely on, at the version in use.

Your SOP rules for validity and seriousness
The MedDRA version in use
Your product dictionary
Reporting rules by market

Prompts

Approved, version-controlled instructions for common safety tasks.

Summarise a case for medical review
Explain a coding suggestion
Draft a follow-up request
Explain why a case is reportable

Why not just give an agent a general model?

A general assistant can read a report, but it does not know your SOPs, the MedDRA version you use or your reporting rules, and it cannot show an inspector why it answered as it did. Through a vGenix MCP server, it uses governed tools that apply your rules and return evidence.

What the agent can never do

Act outside the scope set for it in the control plane
Skip a human gate your SOP requires
Use an unvalidated model or prompt version
Act without leaving an audit trail entry

Governed by vGenixOS

Not a raw AI endpoint

IntaaS is not a language model behind an API key. Every capability runs on vGenixOS, with the same governance that makes the platform suitable for regulated work.

Pharma Control Plane

Every API call and MCP tool call passes the same controls as the platform: identity and scope, AI governance, data protection, human gates, evidence and change control.

Neurosymbolic checks

Neural AI proposes; symbolic AI checks the answer against your rules and the knowledge graph before it is returned, or routes it to a person.

The same kernels

Rules, knowledge, clock and evidence kernels sit behind every service, so an API answer matches what the platform would give.

Evidence kernel

Every call is recorded with its inputs, result, rule, model version and trace ID, ready for audit and inspection.

Versioned and validated

Services are versioned. New versions are released under change control with validation evidence, and you choose when to move. [Confirm version policy.]

Your data, your region

Data stays in the region you choose and is used only as your agreement allows. [Confirm data use policy.]

Ways to use it

Four common starting points

REST API

Keep your safety database

Your intake portal calls classify, check_validity and extract_case_data, then creates a pre-filled case in your existing safety database for your team to review.

MCP

Give your safety copilot real tools

Your internal AI assistant connects to the vGenix case and submission MCP servers. A case processor asks “is this report valid, and when is it due?” and gets a governed answer with evidence.

REST API

Add literature screening to your service

A CRO calls screen_records and identify_cases inside its own literature service, and keeps its client workflows unchanged.

REST API or MCP

Feed your signal analytics

Your analytics team calls detect_signals and build_evidence_pack on its own data cut, and brings the results into its existing review meetings.

Platform or service

Two ways to use vGenixOS

Start with the service and move to the platform later, or use both. They run on the same engine.

Area	vGenix Ai platform	Intelligence as a Service
What you get	Five modules, workflow, workspace and intelligence	Module intelligence as REST APIs and MCP servers
Where your team works	In the vGenix Ai workspace	In your own systems and AI assistants
Workflow and review gates	Built in, configured from your SOPs	Run in your systems; routing signals returned with every answer
Governance	Pharma Control Plane, audit trail, change control	The same, for every call
Best for	Replacing fragmented tools with one platform	Adding intelligence without changing platform

What vGenix Ai provides

Versioned, validated services and MCP servers
Evidence and trace for every call
Release notes and validation evidence for every version
Support and service levels per agreement

What stays with you

Your workflow and the decisions made in it
Acting on routing signals and human gates
Validation of your own integration
Regulatory responsibility for your submissions

Getting started

From first call to production

STEP 1

Choose capabilities

Pick the module capabilities that fit your process.

STEP 2

Sandbox

Call the APIs or connect an agent to MCP servers with sample data.

STEP 3

Configure and validate

Your SOP rules applied; your integration validated with our evidence.

STEP 4

Go live

Production access, monitoring and versioned releases.

Next step

Put vGenix Ai intelligence inside your systems

Tell us which capabilities you need and where. We’ll set up sandbox access and show you the evidence behind every answer.

Book a demo Talk to our team

Recommended

6-week pilot on literature monitoring

Baselines agreed up front
Configured to your SOPs
Measured on your own data
Validation evidence from day one

Start your pilot

Why vGenix

Platform

Solutions

Proof

Resources

Intelligence
as
APIs & MCP servers

Use the intelligence without changing your platform

The same intelligence, unbundled

APIs for systems, MCP for agents

19 capabilities across five modules

Every answer carries its own evidence

Give your AI agents safety expertise, with guardrails

Tools

Resources

Prompts

Why not just give an agent a general model?

What the agent can never do

Not a raw AI endpoint

Pharma Control Plane

Neurosymbolic checks

The same kernels

Evidence kernel

Versioned and validated

Your data, your region

Four common starting points

Keep your safety database

Give your safety copilot real tools

Add literature screening to your service

Feed your signal analytics

Two ways to use vGenixOS

What vGenix Ai provides

What stays with you

From first call to production

Choose capabilities

Sandbox

Configure and validate

Go live

Put vGenix Ai intelligence inside your systems

6-week pilot on literature monitoring

Intelligence as APIs & MCP servers

Use the intelligence without changing your platform

The same intelligence, unbundled

APIs for systems, MCP for agents

19 capabilities across five modules

Every answer carries its own evidence

Give your AI agents safety expertise, with guardrails

Tools

Resources

Prompts

Why not just give an agent a general model?

What the agent can never do

Not a raw AI endpoint

Pharma Control Plane

Neurosymbolic checks

The same kernels

Evidence kernel

Versioned and validated

Your data, your region

Four common starting points

Keep your safety database

Give your safety copilot real tools

Add literature screening to your service

Feed your signal analytics

Two ways to use vGenixOS

What vGenix Ai provides

What stays with you

From first call to production

Choose capabilities

Sandbox

Configure and validate

Go live

Put vGenix Ai intelligence inside your systems

6-week pilot on literature monitoring

Intelligence
as
APIs & MCP servers