Build around
the work.
Start with model access, agent workflows or knowledge. Choose the integration path that fits your application.
In short:An OpenAI-compatible gateway for building AI applications. Use the official OpenAI SDKs or plain HTTP with your Swfte API key.
application → model access → response
Model access
Connect your application to model access. Start by reviewing routing, authentication and usage requirements.
Explore ConnectFollow the published reference.
Read the SDK examples and integration guidance alongside the product you intend to use.
Open the developer referenceBefore your first integration
Agree the credential boundary, test inputs, failure handling and the usage you need to observe.
Discuss an integrationInteractive API Playground
Test API calls in real-time with live code examples and instant results.
curl https://api.swfte.com/agents/v2/gateway/chat/completions \
-H "Authorization: Bearer $SWFTE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "openai:gpt-4o-mini",
"messages": [
{ "role": "user", "content": "Say hello in five words." }
]
}'Your guided integration path
Get up and running in minutes with our step-by-step integration guide.
export SWFTE_API_KEY="sk-swfte-..."Create an API key
Open the Connect console, create an API key and keep it in an environment variable.
# pip install openai
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.swfte.com/agents/v2/gateway",
api_key=os.environ["SWFTE_API_KEY"],
)Point the OpenAI SDK at Swfte
The gateway is OpenAI-compatible. Install the official openai package and set the base URL.
response = client.chat.completions.create(
model="openai:gpt-4o-mini",
messages=[{"role": "user", "content": "Hello, world!"}],
)
print(response.choices[0].message.content)Make your first request
Use a mini model on the free plan. Any OpenAI-compatible client or plain curl works the same way.
Built for developers
Everything you need to build production-grade AI applications.
Complete TypeScript definitions
with autocomplete and compile-time checking.
TypeScript, Python, Go, Ruby
and more with full feature parity.
Token-by-token streaming
with SSE and WebSocket support.
Latency, tokens, costs, errors
tracked automatically.
Frequently Asked Questions
There is no Swfte-branded SDK yet. The gateway is OpenAI-compatible, so the official OpenAI SDKs (Python, TypeScript and others) work when you set the base URL to the Swfte gateway, and so does any HTTP client.
Use API keys stored in environment variables or passed directly to the SDK. We support OAuth2 for enterprise SSO integration and provide secure key rotation.
Free tier: 100 requests/minute. Pro: 1,000 requests/minute. Enterprise: Custom limits with dedicated capacity and burst handling.
The SDK provides typed error classes with detailed messages. All errors include request IDs for debugging and support. We also offer automatic retries with exponential backoff.
Yes! All SDKs support Server-Sent Events (SSE) for real-time streaming. You can receive tokens as they are generated for responsive UIs.