Quickstart

1. Get an API key

Arcware is currently onboarding teams directly. Request access and we’ll set you up with a key — usually within a day. Keys look like sk-arc-… and should live in an environment variable, never in code.

2. Point your SDK at Arcware

Set the base URL and pass your key. That is the entire migration.

from openai import OpenAI
import os

client = OpenAI(
    base_url="https://api.arcware.us/v1",
    api_key=os.environ["ARCWARE_API_KEY"],
)

3. Send your first request

response = client.chat.completions.create(
    model="deepseek/deepseek-v4-pro",
    messages=[{"role": "user", "content": "Summarize this quarter's usage data."}],
)

print(response.choices[0].message.content)

Or from the command line:

curl https://api.arcware.us/v1/chat/completions \
  -H "Authorization: Bearer $ARCWARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4-pro",
    "messages": [{"role": "user", "content": "Hello, Arcware."}]
  }'

4. Stream responses

Pass stream=True and tokens arrive as they are generated — the right default for agents and anything user-facing.

stream = client.chat.completions.create(
    model="deepseek/deepseek-v4-pro",
    messages=[{"role": "user", "content": "Plan a database migration."}],
    stream=True,
)

for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")

Models

One model is live today, with the roadmap close behind. Model IDs are stable — switching models is a one-line change.

Full specs for the live model are on the DeepSeek V4 Pro page, and rates for everything are on the pricing page.

Privacy

Every request is served from U.S.-hosted infrastructure under a Zero Data Retention policy: prompts and generations are not stored after the request completes and are never used for training. This is the default for all customers, not a configuration tier.

Support

Stuck, or need higher limits? Email [email protected] — you’ll get an engineer, not a queue.