Virtual API Keys
Issue scoped credentials with strict model locks and monthly budget caps.
| Project / App | Virtual Key | Allowed Models | Spend / Budget | Status | Actions |
|---|---|---|---|---|---|
| Loading keys... | |||||
Live Activity Stream
Real-time log of every incoming model request, token consumption, and dollar cost.
| Timestamp | Key / Project | Model | Tokens (In / Out) | Cost ($ USD) | Latency | Status |
|---|---|---|---|---|---|---|
| No activity logged yet. | ||||||
Gateway Analytics & Token Metering
Aggregated token throughput and Google Cloud expenditure tracking.
100% Upstream Direct
In: 0 | Out: 0
Direct Google wholesale rates
0 active
Interactive AI Playground
Test prompts, vision payloads, and model behaviors in real-time through your gateway.
Gateway Settings
Developer Integration Manual
Complete guide to connecting Python, Node.js, and HTTP clients to your in-house AI gateway.
Connect using standard official OpenAI client libraries in Node.js or Python simply by changing the baseURL and providing your Virtual API Key.
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "http://localhost:3000/v1",
apiKey: "comp-live-YOUR-KEY-HERE", // Scoped virtual key
});
const res = await client.chat.completions.create({
model: "gemini-3.1-flash-lite",
messages: [{ role: "user", content: "Write a high-converting product headline." }],
max_tokens: 100,
});
console.log(res.choices[0].message.content);from openai import OpenAI
client = OpenAI(
base_url="http://localhost:3000/v1",
api_key="comp-live-YOUR-KEY-HERE"
)
response = client.chat.completions.create(
model="gemini-3.1-flash-lite",
messages=[{"role": "user", "content": "Explain quantum computing in 2 sentences."}]
)
print(response.choices[0].message.content)import OpenAI from "openai";
import fs from "fs";
const client = new OpenAI({
baseURL: "http://localhost:3000/v1",
apiKey: "comp-live-YOUR-KEY-HERE",
});
const base64Data = fs.readFileSync("receipt.jpg").toString("base64");
const res = await client.chat.completions.create({
model: "gemini-3.1-flash-lite",
messages: [
{
role: "user",
content: [
{ type: "text", text: "Extract total amount and store name from this receipt:" },
{
type: "image_url",
image_url: { url: `data:image/jpeg;base64,${base64Data}` }
}
]
}
]
});
console.log(res.choices[0].message.content);Access Google Search Grounding through the native gateway endpoint. Google embeds live web links and fact citations automatically into candidate outputs.
curl http://localhost:3000/v1beta/models/gemini-3.1-flash-lite:generateContent?key=comp-live-YOUR-KEY \
-H "Content-Type: application/json" \
-d '{
"contents": [{"parts": [{"text": "What are the latest breakthrough AI models released this month?"}]}],
"tools": [{"google_search": {}}]
}'🔒 Zero Disk Storage
Images and prompts pass purely through RAM buffers and stream straight to Google. Neither images nor full response payloads are ever written to your gateway server disk.
🛡️ Scoped Model Enforcement
If a developer key attempts to query a model not checked in their key profile (e.g. gemini-2.5-pro), the gateway rejects the call immediately with 403 Forbidden.