Python SDK (beta)

pip install roadie-sdk

export ROADIE_KEY="rd_sk_prod_…"

import os
from roadie import Roadie

roadie = Roadie(api_key=os.environ["ROADIE_KEY"])

res = roadie.chat.create(
    model="openai/gpt-4o",
    messages=[
        {"role": "system", "content": "You are a concise assistant."},
        {"role": "user", "content": "In one sentence, what is Roadie?"},
    ],
    user={"id": "user_42", "plan": "pro"},
)
print(res.text)
print(res.usage, res.cost, res.request_id)

with roadie.chat.stream(model="openai/gpt-4o", messages=messages) as stream:
    for event in stream:
        if event.type == "content_delta":
            print(event["delta"], end="", flush=True)
        elif event.type == "message_end":
            print("\n", event.get("usage"), event.get("cost"))

res = roadie.embeddings.create(model="openai/text-embedding-3-small", input=["hello", "world"])
for row in res.data:
    print(row.index, len(row.embedding))

for model in roadie.models.list().data:
    print(model.id, model.provider, model.capabilities)

token = roadie.client_tokens.create(
    end_user_id="user_42",
    plan="pro",
    ttl_seconds=900,   # the gateway clamps to ≤ 3600
)
print(token.token, token.expires_at, token.end_user_id)

from roadie import (
    Roadie,
    RateLimitError,
    QuotaError,
    BudgetError,
    AuthenticationError,
    RoadieError,
)

try:
    res = roadie.chat.create(model="openai/gpt-4o", messages=messages)
except RateLimitError as err:
    ...  # transient: back off and retry after err.retry_after seconds
except QuotaError:
    ...  # a usage window is exhausted
except BudgetError:
    ...  # the spend budget cap was hit (HTTP 402)
except AuthenticationError:
    ...  # the key is missing, wrong, or revoked
except RoadieError as err:
    print(err.type, err.code, err.request_id)