First response in two minutes
One endpoint, one key, every model. These snippets run as-is once you paste in a key from the API Platform Console — free trial credits are included, no card needed.
1 · cURL
curl https://api.projectares.ai/v1/messages \
-H "x-api-key: sk-ares-YOUR_KEY" \
-H "content-type: application/json" \
-d '{
"model": "claude-sonnet-5",
"max_tokens": 300,
"messages": [{"role": "user", "content": "Say hello in three languages."}]
}'
You get back a message object; the text lives at content[0].text and exact token counts at usage. Free test drive: use model ares-mock — it streams a canned reply and costs nothing.
2 · Python
The platform speaks the standard Messages protocol, so the SDK you already use just works — point it at our base URL:
# pip install anthropic
import anthropic
client = anthropic.Anthropic(
base_url="https://api.projectares.ai",
api_key="sk-ares-YOUR_KEY",
)
msg = client.messages.create(
model="claude-sonnet-5", # or "gpt-5.5", "zai.glm-5", … — same code
max_tokens=300,
messages=[{"role": "user", "content": "Say hello in three languages."}],
)
print(msg.content[0].text)
3 · TypeScript / Node
// npm install @anthropic-ai/sdk
import Anthropic from "@anthropic-ai/sdk";
const client = new Anthropic({
baseURL: "https://api.projectares.ai",
apiKey: process.env.ARES_API_KEY,
});
const msg = await client.messages.create({
model: "claude-sonnet-5",
max_tokens: 300,
messages: [{ role: "user", content: "Say hello in three languages." }],
});
console.log(msg.content[0].type === "text" ? msg.content[0].text : msg);
4 · Switch models by changing a string
Every model on the platform answers on the same endpoint in the same shape. Discover what's available (and what it costs) at runtime:
# the curated picker list (public)
curl https://api.projectares.ai/catalog
# every model + lifecycle metadata (auth)
curl https://api.projectares.ai/v1/models -H "x-api-key: sk-ares-YOUR_KEY"
# live per-model pricing (public)
curl https://api.projectares.ai/pricing
Next steps
- The cookbook — streaming, tool use, prompt caching (big savings), the priority lane, and per-project telemetry.
- API reference — every endpoint, error type, and header.
- OpenAPI 3.1 spec — generate a typed client in any language.