Pirol API.
AI models from the EU with Pirol Online, or on your own hardware with Pirol Local.
— architecture
One line of config.
your_app.py
from openai import OpenAI
client = OpenAI(
base_url="https://endpoint.pirol.ai/v1", # ← only change
api_key="ep_live_…",
)
response = client.chat.completions.create(
model="pirol/pirol-ai", # Pirol AI, no provider key needed
# bring your own key to reach OpenAI or Anthropic models here
messages=[{"role": "user", "content": prompt}],
)— pirol local vs. pirol online
Pirol Local
Local AI, on your own hardware
AI Models
- ✓ Text (open-weight models, on your hardware)
- ✓ Audio Transcription
Platform
- ✓ Kyte Database
- ✓ Background Agents
Pirol Online
AI from the EU, managed by us
AI Models
- ✓ Text (open-weight models, chosen for you)
- ✓ Audio Transcription
Platform
- ✓ Kyte Database (daily backups)
- ✓ Background Agents
Bring your own key — OpenAI, Anthropic and others through the same gateway.
— features
Everything you need, nothing you don't.
Drop-in compatible
Change one URL in your existing OpenAI SDK setup. No new client library, no schema changes.
EU-hosted
Traffic stays on European infrastructure. No transatlantic data transfer for the raw payload.
Full audit log
Every request is logged with its model and token counts — queryable from the console.
Per-workspace keys
Issue scoped API keys per team or product. Revoke instantly without rotating provider credentials.
Picks the model
One model id, the right model underneath. Pirol AI routes each request to the open-weight model that handles it best.
Ready to ship?
Sign up and make your first request in minutes.
Get an API key →> often asked