Skip to main content
Sail is a drop-in replacement for OpenAI-compatible inference providers. It serves the OpenAI Responses (/v1/responses) and Chat Completions (/v1/chat/completions) APIs, so switching from OpenAI — or any OpenAI-compatible provider — is just a base URL, API key, and model name change.

​
1. Get your API key

Sign up at the Sail dashboard and create an API key. Export it where your agent and app can read it: 
export SAIL_API_KEY="YOUR_SAIL_API_KEY"

​
2. See what changes

Already calling the OpenAI Responses API? Then it’s three lines — base URL, key, and a model from the catalog. The request and response are identical:
Sail Responses API
import os

from openai import OpenAI

client = OpenAI(
    base_url="https://api.your-provider.com/v1",  
    base_url="https://api.sailresearch.com/v1",  
    api_key=os.environ["PROVIDER_API_KEY"],  
    api_key=os.environ["SAIL_API_KEY"],  
)

response = client.responses.create(
    model="<your-current-model>",  
    model="<sail-model>",  
    input="Explain the key ideas behind transformers.",
)
print(response.output_text)

​
3. Let your coding agent do it

Across a real codebase the same change touches every call site, env var, and doc reference so let your coding agent run the migration. Open your project and give it this prompt: 
Migrate this codebase from its current LLM provider to Sail.

Reference docs: https://docs.sailresearch.com/llms-full.txt

Sail is drop-in compatible with the OpenAI Responses and Chat Completions
APIs, so keep whichever request shape this code already uses. Make these
changes wherever the LLM client is configured or called:

1. Point the client at Sail:
   - base_url: https://api.sailresearch.com/v1
   - api_key: read from the SAIL_API_KEY environment variable

2. Replace model names with a Sail-supported model. Check
   https://docs.sailresearch.com/models and pick the closest equivalent
   (for example, zai-org/GLM-5.1-FP8).

3. Update env var names and any docs/README references to match.
Review the diff it proposes, then run your tests.

​
Notes

  • Synchronous by default. responses.create blocks and returns the completed response, exactly like OpenAI. Sail is throughput-optimized, so requests can run longer than a typical low-latency API — for very long jobs you can optionally pass background=True to get an ID back immediately and poll, avoiding HTTP timeouts. See the Quickstart.
  • Pick a completion window to trade off cost against turnaround. The default standard window suits most workloads; reach for priority when latency matters or flex for cheap background batches. See Completion windows.

​
Next steps

Quickstart

Make your first request against Sail.

Models

Browse supported models and pick a replacement.

Pricing

Per-token rates by model and completion window.

Support

Email us if you hit anything unexpected.