The OpenAI API alternative where you change one line

UncensoredAI speaks the OpenAI Chat Completions schema — same endpoint shape, same request body, same SSE streaming, same tool_calls. Point your existing SDK at this gateway and every model in the catalog answers on the same wire.

Keep your SDK — openai-python, openai-node, LangChain, LlamaIndex, Vercel AI SDK

Keep your streaming — token-for-token SSE with the same delta format

Cheapest model from $0.04/1M input — no subscription, no seat fee

Migrate in 5 minutes

Before — OpenAI

const client = new OpenAI({
  baseURL: "https://api.openai.com/v1",
  apiKey: process.env.OPENAI_API_KEY,
});

After — UncensoredAI (https://uncensoredai.dev/v1)

const client = new OpenAI({
  baseURL: "https://uncensoredai.dev/v1",
  apiKey: process.env.UNCENSOREDAI_KEY,
});

Verify with curl

curl https://uncensoredai.dev/v1/chat/completions \
  -H "Authorization: Bearer $UNCENSOREDAI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"deepseek-v4.1-flash","messages":[{"role":"user","content":"Hello"}]}'

What changes, exactly

Thing	OpenAI	UncensoredAI
Base URL	api.openai.com/v1	uncensoredai.dev/v1
Auth header	Authorization: Bearer …	Authorization: Bearer … (same)
Endpoint	/chat/completions	/chat/completions (same)
Streaming	SSE, delta.content	SSE, delta.content (same)
Tool calling	tool_calls	tool_calls (same)
Model ids	gpt-4o, …	34 catalog ids — one per page under /models
Billing	prepaid credits	prepaid crypto balance, per token
KYC	account required	none — the key is the account

Price check

Retail per-1M rates on this gateway. The full table with cached-input rates lives on the cost calculator.

Swap the base URL now

Generate a key, change one line, send a request. If anything fails, the error JSON names the fix.

Generate key