For Developers

Add a policy layer to your existing model integration

GuardAI is an OpenAI-compatible API. Change your base URL, keep your existing code, and get policy enforcement on every request.

Try Interactive Demo Request Demo

One line changed. Full protection added.

Swap your base URL. Everything else stays the same.

app.py

1 from openai import OpenAI
2  
3 client = OpenAI(
4−    api_key="your_openai_key"
5+    base_url="https://api.guardai.us/v1",
6+    api_key="your_guardai_key"
7 )
8  
9 response = client.chat.completions.create(
10     model="gpt-5",
11     messages=[{"role": "user", "content": prompt}]
12 )
13  
14 # That's it. Same SDK, same models, same code.

Built for your workflow.

Policy-first enforcement. No vendor lock-in. No headaches.

Optimized for production

Stronger adversarial robustness than built-in model safety. Risk-based routing available for latency-sensitive paths.

Zero vendor lock-in

OpenAI-compatible means you can switch providers anytime. Works with your existing model or any OpenAI-compatible endpoint.

Flexible deployment

Managed cloud, your own VPC, or on-prem containers. Deploy wherever your infrastructure lives, with outbound access to Vertex AI for the defense layer.

Ship in hours, not weeks

No security research required. No training custom models. Just integrate and deploy. Production-ready day one.

Full visibility and control

Comprehensive audit logs, detailed threat analytics, and customizable tuning for your application’s unique requirements.

RAG-ready protection

Prevent chatbots from leaking sensitive data from your knowledge base. Ship RAG applications with confidence knowing your documents are protected.

Compatible with everything.

GuardAI works with your existing tools and workflows out of the box.

Any LLM Provider

Cloud providers, self-hosted models, fine-tuned deployments, or bring your own API. Any OpenAI-compatible endpoint works.

Cloud ModelsOpen-SourceLocal ModelsFine-tuned

Popular Frameworks

Works with LangChain, LlamaIndex, Haystack, or any framework that uses the OpenAI SDK under the hood.

LangChainLlamaIndexCustom RAG

Flexible Deployment

Use our managed cloud API, deploy in your own VPC, or run containers on-prem. All deployments require outbound access to Vertex AI.

Managed CloudYour CloudOn-Prem

Works with what you already use.

Any application that supports an OpenAI-compatible backend. Your own code using the OpenAI SDK. Or on-prem containers for enterprises with custom solutions and data residency requirements.

API INTEGRATION

Managed Proxy

Point any OpenAI-compatible application or SDK at GuardAI. We forward to your model provider with policy enforcement on every request.

Existing apps that support custom backends

Your code using the OpenAI SDK

LangChain, LlamaIndex, and other frameworks

ENTERPRISE

Private Deployment

Run GuardAI containers in your own VPC or on-prem infrastructure. You control network boundaries and data residency.

Full control over your data environment

Custom model configurations

Your infrastructure, your rules

Add GuardAI to your stack.

See the same-base-model benchmark results on the whitepaper.

Try Interactive Demo Contact Sales