Cloud Run
Regional container services for the public, clinic, owner, API, and worker workloads.
Technology & AI
Intelligence for the recovery journey, with clinical authority kept human.
Vetomo combines structured care workflows, deterministic safety controls, and Gemini on Vertex AI to prepare relevant recovery context for veterinary professionals. Google Cloud provides the regional compute, data, identity, storage, messaging, and observability foundation.
Google Cloud-native architectureBuilt on Google Cloud
A production architecture designed for secure regional scale.
The product is designed as a regional, identity-aware SaaS platform on Google Cloud. Each service has a narrow operational responsibility and patient data remains behind authenticated, tenant-scoped boundaries.
Cloud SQL
Authoritative PostgreSQL data with tenant-scoped access controls and audited mutations.
Cloud Storage
Private regional objects for approved media, exports, and platform assets.
Identity Platform & Firebase
Tenant-aware authentication for veterinary teams and secure owner access.
Pub/Sub
Reliable asynchronous delivery with transactional outbox boundaries.
Gemini on Vertex AI
Structured multimodal preparation through workload identity and regional routing.
AI-first product strategy
AI is a foundational intelligence layer—not a bolt-on chatbot.
Vetomo uses AI to turn longitudinal owner updates, structured observations, media, care-plan context, and clinical history into review-ready context. The model never receives authority to diagnose, prescribe, approve, or bypass a deterministic safety rule.
Owner and clinic context
Compose only the evidence needed for the requested workflow.
Structured multimodal understanding
Prepare bounded text, image, and video observations.
Longitudinal preparation
Organize change over time and surface missing or conflicting context.
Professional review
Present traceable preparation while veterinary professionals decide.
Controlled intelligence loop
Context in. Safety checks. Structured preparation. Human decision.
Every AI-assisted workflow follows the same bounded path. Model output is structured evidence for review, never a command to the clinical system.
Compose
Retrieve the minimum authorized episode and evidence context.
Check
Apply deterministic safety rules before any provider call.
Prepare
Gemini returns schema-constrained observations and context.
Re-check
Validate output and apply safety and protocol gates again.
Review
A veterinary professional interprets and decides the next action.
Release status
Implemented deliberately, released through explicit assurance gates.
AI capabilities are implemented behind disabled-by-default feature flags and are activated only after provider, veterinarian, privacy, localization, storage, monitoring, and operational reviews. Ordinary care tasks and account operations continue when an AI provider is unavailable.
Implemented behind controlled feature flags
Gemini is the primary production AI implementation
Vertex AI workload identity is required in production
Human clinical decisions remain mandatory
Product and technology together