How the brain
actually works.
A knowledge graph plus hybrid retrieval, grounded in your documents, deployable from cloud to air-gapped. No black boxes.
Five layers. One platform.
The Certant stack from raw bytes on disk to an answer in someone's hands.
PDFs, Word, slides, scans, images. We parse with structure preserved — tables, sections, headings stay intact. OCR runs on scans.
Your entities — employer, member, contract, vendor — and the relationships between them. Your vocabulary, learned. No manual ontology.
Vector search finds answers by meaning. GraphRAG follows the relationships. Structured query hits your warehouse. We pick what's best per question.
Every answer is grounded. The citation engine attaches the source paragraph. Guardrails refuse rather than hallucinate.
Agent Builder, Analytics, Chatbots. Same retrieval. Same graph. Same citations. The differences are interface, not intelligence.
GraphRAG + hybrid retrieval,
in plain English.
Finds answers by meaning.
Every paragraph in your document gets converted into a numerical fingerprint. When someone asks a question, we fingerprint the question and find paragraphs that mean something similar — not just paragraphs containing the same words.
Vectors shine on open-ended language: "what does our policy say about overtime?" But they break on questions about relationships — and most enterprise questions are about relationships.
Walks the relationships.
Start from a named entity in the graph — a vendor, a contract, a person. Walk the edges. Pull back the connected sub-graph as context. The LLM gets not just the relevant text, but everything connected to it.
Example a vector index can't answer alone: "which contracts does our largest vendor have with our highest-risk member entities?" That's a graph traversal, not a similarity search.
Combined at query time.
For every question, the LLM picks the strategy: vector for prose-heavy lookups, graph traversal for relationship questions, structured query for numbers, often a blend. The path the answer took is observable — you can ask why this answer and see the sub-graph and the cited passages.
How Certant reads documents.
Adaptive chunking
Reads documents the way a person would. Tables stay tables. Sections stay sections. Headings stay headings. We don't slice mid-clause.
Multimodal native
PDFs, Word docs, slides, scans, images. OCR runs where text isn't extractable. Page-image references survive into the citation engine.
Auto-ontology
We learn your vocabulary as we go. "Worker" / "Staff" / "Employee" become one entity, mapped to your schema — no manual setup.
No black boxes.
Every Certant answer traces back through this graph. Pan it. Zoom in. Click anything.
Every node is something in your business. Every edge is a relationship we learned from your data. Edges are typed (e.g. governed_by, employs, requires) and clickable.
Bring your own.
Or use ours.
We don't lock you to a model. Models change every few months. Locking you to one is locking you out of the next breakthrough.
For Cloud, you can pick from AWS Bedrock, Azure AI Foundry, GCP Vertex AI, or our defaults. We route per workload — reasoning models for hard questions, fast models for short answers, embedding models for retrieval.
For Sovereign, you bring your own — open-source weights or a licensed vendor — and run them on your hardware. We support local GPUs (≥80 GB VRAM) and offline embedding models.
From cloud to air-gapped,
in one product.
Cloud
- Hosted in our AU / EU / US regions
- Tenant isolation at the data layer
- SOC 2-aligned controls
- You're live in under 10 minutes
Inside your VPC
- Deploys into your AWS, Azure or GCP account
- Your VPC, your IAM, your KMS keys
- Certant ships the software, you operate the cloud
- Available via cloud marketplaces
Air-gapped
- No outbound connectivity required
- Models, indexes, audit logs all local
- Signed update bundles via removable media
- Designed for IRAP / DEFCON / sovereign workloads
Docker images and Kubernetes manifests are available for any deployment mode. The page you're on isn't a manual — talk to deployment if you need YAML.
Talk to deploymentEvery answer cites its source.
When Certant answers a question, the answer carries the paragraph it came from. Click and you see the original document with the cited passage highlighted. Print it. Send it to your auditor.
There's no "trust us".
What we don't pretend to be.
We're not a foundation model.
We don't train base LLMs. We make the ones you already trust useful on your own data.
We're not a BI suite.
Analytics is one of three surfaces. If all you need is a pivot table, buy a pivot table.
We're not a no-code app builder.
Agent Builder builds workflows. It doesn't build customer-facing apps.
Convinced or sceptical,
talk to us.
We answer technical questions in plain English.