Speed
Find the right data, fast
Teams locate the authoritative dataset and understand it without a hunt through systems and inboxes.
CYRENYX™ · Life Sciences data & AI
Data stays where it lives. Knowledge gets connected.
CYRENYX builds governed layers of knowledge above your existing Life Sciences data estate, connecting clinical, imaging, RWE, omics, research, manufacturing and regulatory environments.
No data migration. No platform replacement.
- Clinical
- Imaging
- RWE
- Omics
- Research
- Manufacturing
- Regulatory
One intelligence layer
It continuously integrates data, metadata, scientific context, lineage, quality and policy across distributed systems, while preserving domain ownership, source-of-truth integrity and data residency.
Governance in every layer
GxP-aligned controls, auditability, lineage, purpose-of-use, privacy and region-specific data requirements apply as data is discovered, accessed and used.
AI-ready, where it belongs
A connected, context-aware, AI-ready data estate: data stays in the systems and regions where it belongs, and becomes discoverable, interoperable, governed and usable across the enterprise.
The problem
Life Sciences data is everywhere. Its context isn't.
Clinical, imaging, real-world, omics, research and regulatory data sit in different systems, owned by different teams. Simple questions about it are still hard to answer:
- Where is the authoritative dataset?
- What does this dataset actually represent?
- Which study, patient population or experiment is it associated with?
- How was it transformed?
- Does it contain sensitive information?
- Can it be reused for this purpose?
- Is it ready for AI?
- Which models, analyses or downstream artifacts depend on it?
How CYRENYX works
From distributed data to trusted scientific intelligence
Five steps run continuously above the platforms you already have. Nothing is migrated: CYRENYX keeps governed references, metadata, relationships and context, and preserves them as data moves across the organisation.
Your data, where it lives
Clinical · Imaging · RWE · Omics · Research · Manufacturing · Regulatory on AWS, Azure, Google Cloud, Databricks, Snowflake and your source systems
- DiscoverFinds data across distributed systems and catalogues its technical and scientific metadata.
- ContextualiseLinks it to studies, protocols, ontology and business meaning, and traces its lineage.
- GovernApplies access, privacy, consent and purpose-of-use controls, with a full audit trail.
- AssessScores quality, flags data issues and evaluates readiness for each AI use case.
- UseOpens governed access for people, applications and AI agents.
Built into every stepPurpose-of-use & access policy
GxP-aligned audit trailLineage & versionsData residency
The outcome
Data your teams can find, trust, reuse and safely put to work with AI
What makes CYRENYX different
A different starting point, not another platform
CYRENYX doesn't replace your data platforms, catalogue tools, agent frameworks or security products. It adds the governed scientific context that connects them. The difference is where it starts.
Where many programmes startWhere CYRENYX starts
Move the data into a new platform, then govern it
Leave the data where it is. Governed references, metadata and context across the systems and regions you already run.
Catalogue technical metadata
Add scientific context. Study, protocol, population and ontology beside the technical metadata.
Document access policies beside the data
Decide every query by policy. Purpose-of-use, consent and access apply to people and AI agents alike.
Start AI with a single model
Start AI with governed context. Specialist agents, each on a model suited to its work, on data assessed for that use.
Observe which model answered
Trace the whole path. From source data, through context, policy and validation, to the AI output.
Data+Scientific context+Governance+AI readiness+Specialist agents
One intelligence layer, built for Life Sciences, above the platforms you already trust.
Inside CYRENYX
What it looks like in use
Screens from the product demos, on synthetic data.
What CYRENYX does
Eight services, and what each one delivers
Each service runs continuously across the systems you already run, not as a one-off documentation project.
ServiceWhat CYRENYX doesOutcome for your teams
Discover
Data discovery
Discovers data assets across clouds, lakehouses, clinical, imaging, research and document systems, in place.
One inventory of every dataset, with the authoritative copy identified.
Discover
Cataloguing
Catalogues technical and scientific metadata: schemas, profiles, owners, studies, endpoints and terms.
People search by what data means, not by where it is stored.
Contextualise
Scientific context
Contextualises data with its study, protocol, patient population, ontology and business meaning.
Anyone can tell what a dataset represents without tracking down the person who built it.
Contextualise
Lineage
Tracks lineage from source, through every transformation, to the analyses, models and submissions that use it.
Changes are made knowing what depends on them, and results can be reproduced.
Govern
Governance
Applies access, privacy, consent and purpose-of-use controls to every query, by people and agents alike.
The right access for the declared purpose, provable from a tamper-evident audit trail.
Assess
Data quality
Assesses quality continuously and identifies data issues as data arrives and changes.
Issues are caught before analysis, submission or model training, not after.
Assess
AI readiness
Evaluates whether data can safely and effectively support each AI use case, and what must be remediated.
Teams know which data is AI Ready, and exactly what to fix when it isn't.
Use
Governed access
Exposes approved data and context to people, applications and AI agents through governed interfaces.
One governed way in for every consumer, so teams reuse data instead of rebuilding it.
The outcome of using CYRENYX
A persistent intelligence layer, and what it changes
Context is no longer lost each time data moves between systems and teams.
Reuse
Reuse instead of rebuild
Existing datasets are found and trusted, so work starts from what already exists.
Compliance
Governed by design
Purpose, consent and access rules are enforced on every query, and every step is auditable.
AI
AI on the right data
AI use cases start from data that has been assessed, documented and certified for that use.
Trust
Context that travels
Study, protocol, quality and lineage context stays with the data wherever it is used next.
Adoption
Value without migration
Start with one domain, workflow or use case on the platforms you run today, and expand from there.
See CYRENYX working
Short demos, on synthetic data
Each demo is captioned on screen, with no sound. Every study, patient and record in them is synthetic.
Demo 01 · 90 seconds
Intelligent Data Discovery
Connect CYRENYX to where a study delivery already lands and watch it identify the data assets, metadata and scientific context: the study, all 172 columns tagged, quality scored and the protocol attached. The pipeline it writes runs on the Airflow you already have.
Demo 02 · preview chapter
Scientific Context and Lineage
See how CYRENYX connects a dataset to its study, protocol and ontology, and traces every column from the EDC through each transformation to the features built on it, including the spreadsheet someone uploaded by hand.
Plays the context & lineage chapter of Demo 03. Ask for the full walkthrough.
Demo 03 · 90 seconds
AI Readiness
Evaluate whether data is suitable for AI. CYRENYX assesses quality, context, lineage, governance and accessibility and highlights what must be remediated before AI can safely use it. Watch a dataset move from Needs Remediation to AI Ready for its intended use.
Live walkthrough
Ask in plain language across clinical, imaging, RWE and omics data, and watch the same policy decide what a person, an application or an AI agent may see.
Demo 04 · live
Governed Search and AI Access
Search across distributed scientific data and expose approved context to applications and AI agents through governed interfaces.
Also · 90 seconds
From a new project to a validated analysis
Governed access for people, end to end. A data manager opens a study project and locks its purpose. Its lead data scientist sets the team's Python and R packages; a missing one is scanned on arrival and validated through a tracked workflow, with a Jira story for every step. Access comes from the access matrix, the containers and Git repository are built, and the team writes its first analysis.
AI readiness
AI readiness is a key part of the platform
Giving an AI model access to enterprise data is easy. Giving it access to the right data, with the right context, quality, lineage and governance is much harder. CYRENYX evaluates whether data is ready to support AI, for the use you intend rather than in the abstract.
Quality
- Data quality
Context
- Metadata completeness
- Scientific context
- Ontology and semantic mapping
Lineage
- Lineage
- Version and freshness
Governance
- PHI / PII sensitivity
- Access policy
- Purpose-of-use
Accessibility
- Machine accessibility
A clear view of every dataset, for every use
- AI ReadyPasses every check for the declared use, and carries a readiness certificate.
- Conditionally ReadyUsable for the declared use under recorded conditions, such as a known limitation or a restricted workspace.
- Needs RemediationBlockers to fix first, each with an owner and a tracked fix; agents handle many of them.
- RestrictedPolicy, consent or sensitivity rules this use out until governance decides otherwise.
AI readiness becomes one measurable outcome of the broader CYRENYX intelligence layer, re-checked at every new data cut. Watch Demo 03.
Governed AI agents
Build AI agents on the same governed context as your data
Most enterprise agent frameworks start with the model and then connect it to tools and data. CYRENYX starts with the governed data context. Because it already connects data, metadata, lineage, ontology, quality, ownership, access policy and purpose-of-use across your estate, agents are built with that context from the beginning.
The agent is not another intelligence layer disconnected from enterprise governance. It operates within the same intelligence layer that understands the data.
A team of specialists, not a single model
Work is planned by a Conductor and delegated to specialist agents. Each one sees only the tools and data its role allows, and can run on the model that suits its work: a general-purpose language model, a domain model, or a validated in-house model with its own model card.
A question or a workflow step“Which DMO-2024-001 subjects had grade 3
pneumonitis, and is that data ready for the safety model?”
ConductorPlans the work and delegates it. Holds no data tools of its own.
- Clinical DataStudy data · SDTM and ADaMe.g. a general-purpose LLM
- ImagingDICOM series · reads · endpointse.g. an imaging model
- Real-world evidenceClaims · EHR · registriese.g. a general-purpose LLM
- Ontology & SemanticMedDRA · LOINC · CDISC termse.g. a terminology model
- Catalog & LineageMetadata · lineage · qualitye.g. a small, fast model
- Therapeutic areaEndpoints · standards · landscapee.g. a validated in-house model
Governance GuardianReviews every answer against policy, consent and purpose before it's released.
Answer or artifactCited to its datasets and studies · every step in the audit trail
From governed data to governed agents
Create and deploy specialised agents for clinical, imaging, RWE, omics, research, regulatory and enterprise workflows, each configured around:
- PurposeWhat the agent is intended to do
- DataWhich governed data it can access
- ToolsWhich APIs and actions it can use
- ModelWhich approved model, or models, suit each workload
- IdentityHow the agent is identified and authorised
- PolicyWhich enterprise controls apply
- ValidationWhat checks occur during the workflow
- Human oversightWhere review or approval is required
Introduce agentic workflows without creating a separate security and governance architecture around every agent.
Security follows the agent
Inside your security architecture
CYRENYX works within your existing security architecture rather than replacing it. Agents operate within your:
- Identity and access controls
- Role-based permissions
- Network boundaries and private endpoints
- Secrets management
- Encryption controls
- Data-residency requirements
- API security
- Logging and monitoring
- Approved model policies
- Regional and organisational data restrictions
Your security controls remain the authority. CYRENYX adds the context, such as ownership, sensitivity and purpose-of-use, to help apply them to AI workflows.
Automated checks
Verification through the lifecycle
Reliability shouldn't depend on trusting a single model response. Depending on the use case, checks include:
- Schema and required-field validation
- Data-quality rules
- Policy and permission checks
- Tool-access checks
- Expected-output validation
- Scientific or business rules
- Source and citation verification
- Confidence thresholds
When a check falls short, the workflow retries, stops, escalates or asks for human review. The aim isn't to claim an agent always knows it's right; it's to make verification a deliberate part of the workflow.
Monitoring
Visibility after deployment
Agent behaviour changes as models, prompts, tools, data and workflows evolve. Teams can monitor:
- Agent activity and workflow execution
- Models used and tools invoked
- Data accessed
- Validation results
- Errors and exceptions
- Latency, resource and model consumption
- Policy events
- Human interventions
An operational record of how agents behave, and which workflows need investigation or improvement.
Traceability
Trace the path from data to AI output
Traditional observability may tell you which model produced a response. CYRENYX connects the broader chain of context, so AI activity can be traced back to the governed data and context that supported it.
- Source data
- Dataset & version
- Scientific context
- Policy & purpose
- Agent
- Model
- Tools
- Validation
- Human review
- Output / artifact
In regulated environments, this record can contribute to the evidence and audit trail that your established validation and quality processes require.
Deploy where the data already lives
Agent workloads and sensitive scientific data don't move to a separate CYRENYX-hosted platform. Depending on your architecture, agents and models run within your existing cloud, Kubernetes or private environments.
Work with the architecture you already trust.
Use the models and tools you approve
A different starting point for enterprise agents
CYRENYX isn't trying to replace every agent-development framework, orchestration platform or security product. Its differentiation is the context underneath the agent.
The same intelligence layer that helps your organisation understand its data answers the questions every agent needs answered first:
- What is this data?
- Where did it come from?
- Can it be trusted?
- Who can use it?
- For what purpose?
Works with what you run
One layer above the platforms you already use
The existing platforms continue to store, process and execute. CYRENYX adds the context, governance and intelligence layer above them.
People, applications & AI agents
ResearchersData stewardsStudy teamsApplicationsAI agentsRegulatory & audit
CYRENYXdiscover · catalogue · contextualise · lineage · quality · governance · AI readiness · governed access
Cloud-native servicesDatabricksSnowflakeAirflowMLflowClinical systemsImaging systemsRWE platformsOmics environmentsAPIs & enterprise services
Your platforms · they keep storing, processing and executing
No migration required
Start with one use case. Expand from there.
CYRENYX does not require you to move your data into another proprietary platform. It maintains governed references, metadata, relationships and context across the systems you already have.
So instead of another migration programme, teams begin with one domain, one workflow or one use case, and expand incrementally.
Your data stays in
Governance
Every query is decided by policy
The same policy engine decides what each person, application and agent sees, as full access, masked, aggregate-only or metadata-only, for the purpose they declared.
Purpose of use
Every session declares a purpose. Change the purpose, and access changes with it, live.
Consent & reuse
Consent bases and reuse restrictions travel with the data, and small cells are suppressed in aggregates.
Tamper-evident audit
Every query, tool call and agent step is written to a hash-chained audit trail that can be verified at any time.
Secrets by reference
Credentials never land in the catalogue, the logs or the audit trail, and agents can read but never change data.
Live demo
Try it on synthetic data
The live demo runs Demo Pharma, a fictional sponsor with oncology, immunology and neuroscience studies. Sign in as a scientist, a data owner or a steward and watch access change with the role and the purpose.
- Sponsor
- Demo Pharma (fictional)
- Studies
- DMO-2023-014, DMO-2024-001 and more
- Data
- Clinical, imaging, omics, registry and documents, all synthetic
- Access
- By invitation; personas from scientist to data owner
Customisable
Fitted to your architecture, not the other way round
No two Life Sciences landscapes look alike. Every CYRENYX deployment is customised to the client's own architecture: where it runs, what it connects to, who signs in and which rules apply. See the reference architectures.
Deployment
Runs where you choose
In your own AWS, Azure or Google Cloud account, or on your Kubernetes cluster, set up with infrastructure as code.
Integration
Connects to your platforms
Over 25 ready connectors, from Veeva Vault, Medidata Rave and SharePoint to DICOMweb, FHIR and S3, plus new ones for your own systems.
Identity
Your sign-in
Single sign-on with Microsoft Entra ID, AD FS, Okta or any SAML 2.0 provider, with your groups mapped to roles.
Governance
Your rules
Purposes of use, access policies, consent bases and approval workflows modelled on how your organisation already works.
Organisation
Your structure
Affiliates, CROs and partners as separate tenants, each with its own data residency, encryption keys and licensed modules.
Context & AI
Your ontology and models
Extend the ontology with your own entities and terms, and use Claude through your own account or run without an external model.
Reference architectures
CYRENYX inside your cloud: AWS, Azure or Google Cloud
The same landing zone on every cloud, built as code in your own organisation: a private Kubernetes cluster in a private network, your keys, your logs, and your data platforms read where they are. Pick a cloud to see its services.
Across every cloud you use
Most Life Sciences companies run more than one cloud. CYRENYX runs in the one you choose and governs data that stays where it lives, wherever that is.
CYRENYX
Keep the data where it is. Add the intelligence layer above it.
Discover. Contextualise. Govern. Assess. Use.
And know when your data is ready for AI. Talk to Astatine Data Services about a walkthrough for your team, or a pilot fitted to your own architecture.