Quickstart
In one breath: point the OpenAI SDK you already use at a new base URL, and every response comes back wearing a receipt.
1. Get a key
Keys are issued through your workspace admin or by talking to us. If your deployment is in-perimeter, your endpoint URL is your own; the examples below use the hosted one.
2. Make your first call
from openai import OpenAI
client = OpenAI(
base_url="https://infer.zkagi.ai/v1",
api_key="zk-...",
)
resp = client.chat.completions.create(
model="zk-default",
messages=[{"role": "user", "content": "Summarize this clinical note: ..."}],
extra_body={"tier": "private"}, # foothill | private | sovereign
)
print(resp.choices[0].message.content)
Same shape in curl:
curl -s https://infer.zkagi.ai/v1/chat/completions \
-H "Authorization: Bearer $ZKAGI_KEY" \
-d '{"model":"zk-default","tier":"private","messages":[{"role":"user","content":"..."}]}'
3. Meet your receipt
Every response carries one:
"receipt": {
"trust_domain": "mathematics",
"tier": "private",
"egress": "none",
"proof": "stark:0x9f3a"
}
Keep it. It's the compliance artifact your future self hands to an auditor.
4. Go deeper
- Ask questions of your own documents with a proof as the key
- Turn the tier up to sovereign for your batch workloads
- Browse the full Zynapse API reference for proofs, media generation and key management
Where this stands
The hosted endpoint, tiers Standard and Private, and receipts are live. The
sovereign tier value routes to the MPC path, which is active R&D and
enabled per engagement. Endpoint names and fields shown here can evolve while
we're pre-GA; the Status board tracks it.