Quickstart
Add Autrace to your AI application in under 5 minutes. No servers to run, nothing to install — just change one URL.
Step 1 — Create an account
Go to app.autraceai.com and sign up free. Your API key is generated automatically during onboarding — it looks like aut_live_xxxxxxxxxxxxxxxx.
Step 2 — Change one line in your code
Replace your LLM provider's base URL with the Autrace gateway. Your existing API key stays the same — Autrace forwards your requests to the model provider.
- Python (openai SDK)
- Node.js (openai SDK)
- cURL
from openai import OpenAI
client = OpenAI(
api_key="aut_live_YOUR_KEY", # your Autrace key from Dashboard → API Keys
base_url="https://gateway.autraceai.com/v1"
)
response = client.chat.completions.create(
model="openai/gpt-5.5",
messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)
import OpenAI from 'openai';
const client = new OpenAI({
apiKey: 'aut_live_YOUR_KEY', // your Autrace key from Dashboard → API Keys
baseURL: 'https://gateway.autraceai.com/v1',
});
const response = await client.chat.completions.create({
model: 'openai/gpt-5.5',
messages: [{ role: 'user', content: 'Hello!' }],
});
console.log(response.choices[0].message.content);
curl https://gateway.autraceai.com/v1/chat/completions \
-H "Authorization: Bearer aut_live_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "openai/gpt-5.5",
"messages": [{"role": "user", "content": "Hello!"}]
}'
Step 3 — Check your dashboard
Open Dashboard → Logs.
Every request is:
- ✅ Logged with full request/response, model, tokens, latency
- ✅ PII scanned — SSNs, emails, credit cards redacted before reaching the model
- ✅ Hash-sealed — tamper-evident audit trail
- ✅ Rule-checked — any policy rules you set are enforced
That's it. Your application is now protected.
How it works
Your App → Autrace Gateway → Model Provider (OpenAI, Anthropic, etc.)
↓
PII Filter
Rule Engine
Audit Log
Usage Tracking
Autrace sits between your application and the AI model. You get full visibility and control without changing how your code calls the API.
Try PII filtering
Send a request containing sensitive data:
curl https://gateway.autraceai.com/v1/chat/completions \
-H "Authorization: Bearer aut_live_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "openai/gpt-5.5",
"messages": [{"role": "user", "content": "My SSN is 123-45-6789, help me with my taxes."}]
}'
In Dashboard → Logs, the prompt shows My SSN is [SSN_REDACTED], help me with my taxes.
The model never received the real SSN.
Next steps
- Set up policy rules — block prompt injection, keyword filters, rate limits
- View API reference — full endpoint documentation
- Connect integrations — LangChain, LlamaIndex, and more
- Manage API keys — rotate, scope, and revoke keys