Analyze anything.
Understand everything connected to it.
GeNexus AI is the Relational Intelligence Discovery Engine, a living knowledge graph that turns any question, concept, or file into an interactive 3D map of everything connected to it, enriched with real context, embedded media, and deep relational structure.
GeNexus returns more than a list of facts. It provides much valued context in a living, navigable graph of everything that matters about your topic: entities, relationships, and media, all at once.
Research focused on building context and discovering relationships.
Every idea exists in a web of connections. When you understand only the facts, you understand a fraction of the truth. GeNexus maps the full relational structure: causes, effects, key figures, institutions, timelines, and mechanisms, rendered as a navigable 3D graph in under a minute.
Query any topic and GeNexus dispatches an AI enriched with verified source context, returning a rich set of interconnected nodes with descriptions, properties, and embedded media, all in a single pass.
Most AI tools rely purely on training data. GeNexus grounds every discovery in verified knowledge, drawing from verified external sources before the AI writes a single word.
The result is output that reflects reality. Facts are drawn from verified sources. Relationships map what actually happened, not what a model approximates from training.
When no verified source data is found for a topic, GeNexus detects this automatically and switches to uncertainty-aware extraction.
01
Pattern detection
Before any AI call, GeNexus scans the query for recency signals to determine the knowledge strategy: recent events, newly discovered vulnerabilities, and emerging topics each trigger a different approach.
02
Verified source fetch
The topic is queried against authoritative knowledge sources. The best-matching content, including summaries, structured facts, and context, is retrieved and prepared as injection context for the AI.
03
Context-enriched extraction
The AI receives the query plus the verified context block. It builds the knowledge graph grounded in verified facts, not statistical recall.
04
Source attribution
When verified source data was found, every node surfaces a source link for that entity. When it wasn't, no link is shown. No false confidence.
03 · Capabilities
One engine. Many entry points.
Topic discovery is the fastest path, but GeNexus reads anything: files, codebases, infrastructure plans, legal documents. Every entry point produces the same living, navigable graph.
Core
Topic Discovery Search
The fastest knowledge experience GeNexus offers. Query any topic, a medical condition, a historical event, a technology, a legal case, and receive a rich knowledge graph in moments. The AI works from verified source context, generating node descriptions, relational edges, and media queries simultaneously, using an embedded schema that preserves relational density even under token constraints.
Verified sourcesMedia per nodeRelational edges
Expand
AI Deep Discovery
Select any node and trigger AI Discovery to dive deeper. GeNexus expands the graph with new nodes and relationships branching from that entity alone, letting you explore any connection as its own sub-universe.
Node expansionMerge & deduplicate
AI
Universal File Extraction
Extended file support(PDF, Word, Excel, PowerPoint, image, Terraform, SQL, YAML, JSON, even zip files) and GeNexus extracts a structured relationship graph. Domain-specific prompts surface the right entities for each file type automatically.
Vision-capableBinary extraction
Render
3D Force-Directed Graph
Every discovery renders as a real-time force-directed 3D graph. Nodes are color-coded by type, draggable, zoomable, and camera-flythrough navigable. Select any node for its full info panel: description, properties, relationships, and embedded media.
Switch to the D3-powered 2D view for a collision-free SVG layout. Search nodes, hover for details, export to PDF. Ideal for stakeholder reports, architecture reviews, and research documentation.
D3 force layoutPDF export
Projects
Project Architecture Breakdown
In the age of AI, understanding your codebase is paramount. Point GeNexus at a codebase, via GitHub import or direct file upload, and receive a full architectural graph: modules, services, dependencies, interfaces, and data flows. Security evaluation, tech stack analysis, and design document generation included.
Research that shows its sources and plays its evidence.
Many nodes in a GeNexus graph are a doorway to relevant video content. YouTube videos are embedded directly where available. Click a node, open the info panel, and watch the relevant documentary, lecture, or highlight clip without leaving the graph.
When a confirmed video ID isn't available, GeNexus generates a precision YouTube search query scoped to that exact node, like "dopamine pathway explained" rather than "Parkinson's disease videos." Wikipedia thumbnail images surface automatically for geographic, biographical, and scientific nodes.
Source links appear only when verified. A source link means the system retrieved real data for that node. No link means the AI worked from its own knowledge, and the absence tells you something.
Photosynthesis · embedded video, playing in-panel
▶
Embedded YouTube, per node
The AI identifies known video IDs from training (documentaries, lectures, highlights, explainers) and embeds them directly. Up to 3 videos per node, playable inside the graph panel.
⌕
Precision search queries as fallback
No confirmed video? GeNexus generates a node-specific YouTube search query scoped to the right channel, format, and content type. Domain-aware: academic topics route to Khan Academy and TED-Ed; sports to highlights.
📷
Wikipedia thumbnail images
Post-render background enrichment fetches Wikipedia images for biographical, geographic, and scientific nodes, rendering them in the info panel alongside the description.
05 · How it works
From question to graph, in four moves.
Context first, extraction second, structure third, render fourth. Every discovery takes the same path, whether the input is a sentence or a SQL schema.
01
Ground
Before the AI sees your query, GeNexus retrieves verified source context, including article summaries, structured facts, and relevant content, and prepares it as injection context.
02
Extract
The AI receives your query plus the verified context. Domain-aware prompts instruct it to produce entities, relational edges, media queries, and properties simultaneously, in one pass.
03
Structure
Output is parsed into a structured schema: nodes with properties and outgoing relationship arrays, flattened into a deduplicated link graph with color and metadata intact.
04
Enrich & Render
The 3D graph renders immediately. In the background, YouTube lookups and Wikipedia image fetches run in parallel, and nodes light up with media as they resolve, without blocking the graph.
06 · Use Cases
Built for context based learning focused on relational connections.
Research, engineering, law, finance: the domain changes. Everything is relational. GeNexus makes those relations visible.
Research & Learning
Concepts that map themselves.
Students, researchers, journalists, and analysts use GeNexus to build instant mental models of complex topics. Query a subject and receive its full relational landscape: key figures, sub-fields, historical context, mechanisms, and applications, each with embedded explainer videos and verified citations.
Engineering teams use GeNexus to map codebases, infrastructure plans, and system dependencies. Import from GitHub for a full architectural breakdown with High and Low Level Design report generation, or drop a Terraform plan to watch resource dependencies render as an animated, navigable graph.
Legal teams extract parties, motions, evidence chains, and precedent relationships from case briefs. Financial analysts surface account flows, offshore structures, and intercompany transfers from ledgers and disclosures. Complex document relationships that take hours to map manually render in seconds.
Party identificationEvidence linkageCapital flow mappingRisk relationships
07 · Providers
Bring your own intelligence.
Claude Sonnet, OpenAI GPT-4o, and Google Gemini are interchangeable extraction backends. API keys live in your browser: no server, no proxy, no lock-in. Each provider uses its native file handling and model strengths.
No API key? GeNexus includes a built-in Haiku tier for instant topic discovery. No setup required.
Anthropic · Claude Sonnet
Native PDF blocks · deepest relational extraction
recommended
OpenAI · GPT-4o
Client-extracted text · vision-capable
active
Google · Gemini Pro
Inline multimodal handling
active
Built-in · Claude Haiku
No key required · topic discovery tier
free tier
08 · File Formats
Extensive file format compatibility.
Structured formats parse natively in the browser. Binary documents extract client-side. Images pass to vision models. Unknown formats fall back to plaintext; the AI infers intent and builds the graph regardless.
No upload server. No third-party processor. Your files stay in your tab.
.jsonnative
.tfterraform
.sqlschema
.yamlconfig
.mdmarkdown
.pdfextract
.docxmammoth
.xlsxsheetjs
.pptxjszip
09 · Pricing
Start free. Upgrade when you scale.
Every plan includes the full discovery experience: topic search, file extraction, 3D & 2D views, and media-enriched graphs. Pro unlocks the Claude Sonnet engine, higher limits, and cloud history.
Free
$0
forever · no card required
Everything you need to explore. Powered by the built-in Claude Haiku engine.
Two ways to go premium: Pro ($9.99/mo) runs on our managed Claude Sonnet — zero setup, plus cloud sync. BYOK ($99 once) runs on your own provider key: pay once, no subscription, with no usage markup. Both unlock the enhanced Design & Security diagram suite; cloud sync stays exclusive to Pro.
Start your first discovery.
Query a topic. Drop a file. Watch everything connected to it emerge in three dimensions. No install, no setup. Just your browser, your curiosity, and the graph.
