AI GatewayOne SDK. Any model, any provider.
Twelve OpenAI endpoints, native Anthropic, native Gemini. Swap SDKs, swap providers, swap models — without editing your app.
02 — Line migration
Keep your SDK, Change the base URL.
import openaiclient = openai.OpenAI(api_key="vsk_...", # Verosek virtual keybase_url="http://your-gateway/v1", # ← only line that changed)response = client.chat.completions.create(model="gpt-4o",messages=[{"role": "user", "content": "How many users are in the database?"}],)# Tools handled internally. Policy enforced. Every step audited.print(response.choices[0].message.content)
Run Claude through the OpenAI SDK. GPT through Anthropic. Any combination.
Fifteen endpoints. Three SDKs. One gateway.
<30 ms P99 overhead
Frequently Asked Questions
A proxy forwards the call and checks a key. Verosek governs the call itself: what runs, on what data, under what limits, before anything executes. Routing is the easy part. Governing the action is the hard part, and that is the part we own.
One line, your base_url. Your existing OpenAI, Anthropic, and Gemini code keeps working. You get full governance without a rewrite.
Every action is checked against your policy before it runs, down to the specific operation and the data it touches. Disallowed actions never reach your systems. Most tools can only grant or revoke a key. Verosek decides per action.
No. The overhead stays invisible in production, and a slow or failing provider is routed around for you automatically.
