rxMaps
Semantic knowledge infrastructure
Transforms unstructured enterprise documents and datasets into persistent, exportable semantic reasoning maps that capture evidence, relationships, decisions and contradictions with full provenance.
SourceLatent-Sense Technologies
AIVancouver, BC
Vancouver-based AI R&D lab building a model-agnostic reasoning and knowledge-graph layer — rxMaps, ReX and rxOrchestrator — that adds auditable, evidence-first reasoning on top of an enterprise's existing LLM or AI stack.
Canadian innovation builds a brighter tomorrow.
About
Source: latentsense.aiLatent-Sense Technologies (LST) is a core AI R&D lab headquartered in Canada. We turn cutting-edge AI research in inference, interpretability and neuro-symbolic reasoning into deployable products for enterprises and public sector agencies. We specialize in glass-box AI reasoning. All of our technologies are glass-box, built in-house as core AI systems that can run as a standalone ecosystem or plug into existing stacks.
Named to Canada's Top 100 AI Startups list for 2026 by the ALL IN AI conference. allinevent.ai
Describes itself as 'a core AI R&D lab headquartered in Canada' whose reasoning technologies are 'built in-house as core AI systems,' not resold third-party models. latentsense.ai
Publishes a documented REST API (rx-map creation, relationship discovery, PII redaction with a tunable relevance cutoff, and a stateful ReasonerX reasoning-graph chat endpoint). controller.latentsense.com
Ships an open-source Python SDK on GitHub for its rxMaps, PII-redaction and relationship-discovery API endpoints. github.com
Listed on AWS Marketplace as 'Cognitive AI and Semantic Infrastructure For Auditable AI' under seller Latent-Sense Technologies Inc. aws.amazon.com
Products & Services
Company site: latentsense.aiProducts Latent-Sense Technologies publishes on its own site. Each card links to the page it was taken from.
ReX
Evidence-first reasoning engine
A neuro-symbolic reasoning layer that detects contradictions, builds causal chains and enforces policy over any LLM's outputs, turning it into a structured, auditable reasoner via a persistent reasoning graph.
SourcerxOrchestrator
Agent orchestration (beta)
Routes decisions, actions, escalations and human-in-the-loop steps across specialist agents under shared context and policy constraints, logging every step into a unified auditable evidence trail.
SourceReDiD
PII de-identification
Redacts personally identifiable information from documents (txt, csv, json, pdf, html) with a tunable relevance-cutoff parameter controlling redaction strength.
SourceDeployment & Services
Company site: latentsense.aiHow Latent-Sense Technologies says its technology can be deployed and delivered.
API and SDK access
Serverless, API-first reasoning infrastructure available via a documented REST API, an open-source Python SDK, and MCP for enterprise integration without re-platforming or retraining existing models.
AWS Marketplace deployment
Deployable directly from AWS Marketplace as a packaged offering; other cloud marketplaces are stated as coming soon.
Technology Focus
Company site: latentsense.aiTerms the company uses about its own work.
Neuro-symbolic reasoningKnowledge graphsEvidence-first AIPII de-identificationExplainable AI
Similar Companies
Other companies in AI.
