Truefoundry GCP

Trusted by enterprises

Powering production AI at scale

Fortune 500s across healthcare, financial services, and technology trust TrueFoundry on GCP.

GCP Marketplace

Find us where you already buy software

Available directly on GCP Marketplace. Subscribe and deploy without opening new vendor relationships or going through separate procurement.

Google Cloud Commitments

Your GCP commitment works here too

TrueFoundry purchases through GCP Marketplace count toward your existing Google Cloud Commitments. Spend budget you've already allocated.

Purpose built
Purpose-built for Vertex AI. Not just compatible with it.TrueFoundry is purpose-built for GCP. That means your Vertex AI models, your Guardrail policies, your GKE clusters, and your IAM roles all work exactly as they are — with a unified control plane sitting above them that your whole organization can use safely.

Better together

What Vertex AI and TrueFoundry unlock together

Three native integrations built around how GCP enterprises actually work.

Vertex AI's model power, available to
your entire organization

Vertex AI + TrueFoundry AI Gateway

GCP Vertex AI gives you access to the world's best foundation models. TrueFoundry makes that access available to every team in your organization — with one unified API, per-team usage controls, and full cost visibility.

One OpenAI-compatible endpoint for all Vertex AI models
Per-team RBAC, token budgets and usage
logs
Service Account, Workload Identity, and API Key auth supported

The safety policies you built in GCP, enforced across every team

GCP Model Armor Guardrails + TrueFoundry

GCP Model Armor Guardrails lets you define the safety and compliance policies your organization needs. TrueFoundry enforces those same policies across every model request, every team, and every application — automatically.

PII detection and redaction inline on every request
Content filtering using your existing guardrail configs
Validate and mutate modes — block or anonymize

Your self-hosted models and
internal tools, unified in one place

GCP GKE + TrueFoundry AI Gateway

GCP GKE gives you the infrastructure to run your own models and internal services alongside Vertex AI. TrueFoundry's AI Gateway combining LLM, MCP, and Agent Gateway — brings them all under one roof.

Self-hosted models on GKE alongside Vertex AI, one endpoint
Internal MCP servers on GKE, governed centrally
Public MCPs accessible through the same gateway

One gateway for every model and tool your teams need

Purple gradient square with white background, shiny surface, and rounded corners in rhombus shape.

Connect

Whether your teams are using Vertex AI models or self-hosted models running on your own GKE cluster, TrueFoundry connects everything through one unified API. The same goes for tools — register your internal MCP servers alongside publicly available ones and give your agents governed access to all of them from one place.

Vertex AI models
Claude, Llama and every model GCP makes available.
Self-hosted models on GKE
Your own models, on your own infrastructure
Self-hosted MCP servers on GKE
Your internal tools, governed centrally
Public MCPs
The broader MCP ecosystem, through the same gateway

Deploy new model catalogue with Llama and other AI models for text generation tasks available.

Your policies, enforced consistently across every team and every model

Govern

Governance shouldn't be something you bolt on after the fact. TrueFoundry lets you enforce safety policies, access controls, and cost limits across every model request — whether it's coming from Vertex AI or a self-hosted model on GKE.

Google Model Armor Guardrails
Your existing Google safety policies, enforced at the
gateway
Custom guardrails
Build and enforce policies tailored to your organization
Third-party guardrails
Bring in providers you already trust
RBAC
Control which teams and users can access which models

Guardrails configuration code with OpenAI and Anthropic models, PII data protection and content filters.

Complete visibility into how your organization uses Vertex AI

Observe

GCP gives you powerful AI infrastructure. TrueFoundry gives you the visibility layer on top of it — so your platform team always knows how Vertex AI and your self-hosted models are being used, by whom, at what cost, and how they're performing. Together, nothing runs blind.

Cost tracking by team
See exactly how Vertex AI spend breaks down across
teams
Token usage
Track consumption across every model and every user
Latency
Monitor performance across Vertex AI and self-hosted
models
Full audit trail
A complete record of every request for compliance and
security teams

Token usage and request metrics dashboard with graphs and charts displaying input output data analysis.

Getting started

Up and running in three steps

No rip-and-replace. TrueFoundry sits above your existing GCP stack without changing any application code.

Deploy in your GCP
account

Install via Helm on your existing GKE cluster or get started directly through GCP Marketplace. Either way you're live inside your own VPC in under an hour, no new infrastructure, no new cloud accounts.

Connect your Vertex AI models and policies

Point TrueFoundry at your GCP account, select your Vertex AI models, and wire in your existing Guardrails configurations. Your security policies are enforced from day one.

Give your teams one endpoint for everything

From that point on, every team in your organization uses one unified API for Vertex AI models, self-hosted models on GKE, and internal tools — with full visibility and cost controls for the platform team.