OpenAI-compatible inference
One endpoint. Explicit model access.
Use the OpenAI client you already have. Each access key can be scoped independently by model, budget, request rate and token rate.
Endpoint
https://api.tensorapi.cloud/v1
Standard chat completions and SSE streaming. No custom SDK required.
Live models
- Loading catalog…
curl
curl https://api.tensorapi.cloud/v1/chat/completions \
-H "Authorization: Bearer $TENSOR_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"qwen3.8-max","messages":[{"role":"user","content":"Hello"}]}'
Python
from openai import OpenAI
import os
client = OpenAI(
base_url="https://api.tensorapi.cloud/v1",
api_key=os.environ["TENSOR_KEY"],
)
r = client.chat.completions.create(
model="qwen3.8-max",
messages=[{"role":"user","content":"Hello"}],
)
print(r.choices[0].message.content)
Compatibility
Use /chat/completions and /models under the base URL. Set stream: true normally; the gateway is configured not to buffer SSE.
Access controls
Keys can be revoked immediately and constrained to models, spend ceilings, RPM, TPM, concurrency and expiry. Prompt and completion bodies are not retained in usage logs.