Docs · v0
Build with Vinci.
A minimal quickstart. Full developer documentation arrives in September 2026, after the August 8 launch.
API quickstart
Call the hosted endpoint.
Point any OpenAI-compatible client at try.getsimpledirect.com once vinci-studio is live.
curl https://try.getsimpledirect.com/v1/chat/completions \
-H "Authorization: Bearer $SIMPLEDIRECT_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "vinci-studio",
"messages": [{ "role": "user", "content": "Summarize OSFI E-23." }]
}'from openai import OpenAI
client = OpenAI(
base_url="https://try.getsimpledirect.com/v1",
api_key=os.environ["SIMPLEDIRECT_KEY"],
)
resp = client.chat.completions.create(
model="vinci-studio",
messages=[{"role": "user", "content": "Summarize OSFI E-23."}],
)
print(resp.choices[0].message.content)Self-host
Run the open weights.
Download vinci-studio from Hugging Face and serve it with vLLM or transformers. Zero Data Retention is automatic — the model runs entirely on your infrastructure.
View on Hugging FaceMore lands in September.
RAG integration, SDKs, and the full reference arrive after launch. Source on GitHub.
APIs that guess vs cite
An API that shows its sources.
Returns a citation you have to verify yourself.
Cites the source it used — or returns 'I don't know'.
Your prompts may train the model.
Zero Data Retention by default; your payloads aren't kept.
Rewrite everything to leave.
Apache 2.0 weights behind the same API — self-host and keep your code.
Launching August 8, 2026
Follow the build.
Open weights, a public Constitution, and a model you can verify — the day it ships. Get the launch in your inbox. No spam, just the drop.
or follow along — Follow on X