Kryyt / Developers
Use an OpenAI-Compatible Client
Reuse the Chat Completions interface, not assumptions about every OpenAI endpoint.
Change your connection
Python
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.kryyt.com/v1",
api_key=os.environ["KRYYT_API_KEY"],
timeout=600,
max_retries=0,
)Node.js
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.kryyt.com/v1",
apiKey: process.env.KRYYT_API_KEY,
timeout: 600_000,
maxRetries: 0,
});Use a Kryyt project key, not an OpenAI key or your Google sign-in token. Set the model to deepseek-v4.1-flash, or discover available IDs through GET /v1/models.
Audit before switching
| Area | Kryyt behavior |
|---|---|
| Authentication | Kryyt project Bearer key; purchased credits only. |
| Models | Current active IDs from /v1/models. Roadmap models are not requestable. |
| Streaming | SSE, optional text deltas, and separate usage chunks. |
| Developer role | Normalized to system before forwarding. |
| Token bounds | max_completion_tokens is mapped to max_tokens. |
| n | Only one choice is supported. |
| Tool execution | Your application executes declared functions. |
| Responses / Assistants / embeddings / audio / image generation / file uploads | Not exposed by this gateway. |
| store / metadata / user / service_tier | Ignored; do not rely on OpenAI-specific persistence or routing behavior. |
| Pricing and quotas | Kryyt account rules apply. OpenAI account balances and allowances do not transfer. |
Validate your integration
- List models with your project key.
- Make one small non-streaming request.
- Test SSE and an interrupted stream.
- Test your required image inputs and function-call loop.
- Handle insufficient credit separately from rate limits.
- Review usage in the Kryyt account before increasing traffic.