PROVIDERS

ONE GATEWAY, EVERY AI PROVIDER

Connect to all major providers and 50+ models through a single integration. Direct API, cloud gateways, enterprise platforms, and self-hosted, all with unified security and zero vendor lock-in.

Key Capabilities

Zero Lock-in

Switch providers with a config change. Compare performance and cost across vendors in real-time.

Bring Your Own Keys (BYOK)

Keep your negotiated enterprise pricing. Store keys encrypted, rotate without code changes.

Platform-Managed

Use PromptGuard-managed credentials. No API keys to manage or rotate yourself.

Self-Hosted & Local

Connect to local Ollama instances, vLLM deployments, or other self-hosted models. Complete data control and privacy.

Unified Security Layer

Every provider, every model protected by the same threat detection, PII scanning, and jailbreak prevention.

Agent Framework Support

First-class integration with OpenClaw, LangChain, LangGraph, CrewAI, Pydantic AI, OpenAI Agents SDK, and Vercel AI SDK.

Cloud IAM Support

Use GCP Service Account, AWS IAM, AWS Bedrock, or Azure AD credentials for enterprise SSO and secure access.

How It Works

1

Connect

Configure your API keys or connect your cloud accounts securely in the PromptGuard dashboard.

2

Update Endpoint

Change one line of code in your application to point to the PromptGuard proxy URL.

3

Route & Protect

We automatically route requests to the best provider, scan for threats, and cache responses.

Switching Providers is Trivial

typescript
import OpenAI from 'openai';

// 1. Point to PromptGuard
const client = new OpenAI({
  baseURL: 'https://api.promptguard.co/api/v1',
  apiKey: process.env.OPENAI_API_KEY, // Or any other provider key
  defaultHeaders: {
    'X-API-Key': process.env.PROMPTGUARD_API_KEY
  }
});

// 2. Just specify the model you want
const response = await client.chat.completions.create({
  // Want Anthropic instead? Just change the model name!
  // model: "claude-3-opus-20240229",
  model: "gpt-5-nano",
  messages: [{ role: 'user', content: 'Hello!' }]
});

// Security, caching, and observability are automatically applied.

Why PromptGuard for Model Routing?

PromptGuard
  • Unified API for 50+ models
  • Built-in Enterprise Security
  • Semantic caching for cost savings
  • Granular per-provider analytics
  • Automatic failover routing
Other Solutions
  • Vendor-specific APIs
  • Separate security per provider
  • No cross-provider caching
  • Fragmented observability
  • Manual handling of provider outages

Ready to connect your models?

Start routing to any AI provider today with built-in security, caching, and observability.