ᛞ  Drop-in migration

Change one line.
Run on the open network.

OpenTau speaks the OpenAI API exactly. Keep your code, your SDK, your prompts — just point the base URL at OpenTau and every request runs on the decentralized, open network instead of one company's servers.

  the only diff
from openai import OpenAI client = OpenAI( api_key=OPENAI_API_KEY, api_key="otau_your_key_here", base_url="https://api.opentau.ai/v1", ) # everything below is unchanged resp = client.chat.completions.create( model="meta-llama/llama-3.3-70b-instruct", messages=[{"role": "user", "content": "Hello from the open network"}], ) print(resp.choices[0].message.content)
from openai import OpenAI client = OpenAI( api_key="otau_...", base_url="https://api.opentau.ai/v1", # ← the one line ) resp = client.chat.completions.create( model="meta-llama/llama-3.3-70b-instruct", messages=[{"role": "user", "content": "Hi"}], ) print(resp.choices[0].message.content)
import OpenAI from "openai"; const client = new OpenAI({ apiKey: "otau_...", baseURL: "https://api.opentau.ai/v1", // ← the one line }); const resp = await client.chat.completions.create({ model: "meta-llama/llama-3.3-70b-instruct", messages: [{ role: "user", content: "Hi" }], }); console.log(resp.choices[0].message.content);
curl https://api.opentau.ai/v1/chat/completions \ -H "Authorization: Bearer otau_..." \ -H "Content-Type: application/json" \ -d '{ "model": "meta-llama/llama-3.3-70b-instruct", "messages": [{"role":"user","content":"Hi"}] }'
# pip install opentau from opentau import OpenTau client = OpenTau(api_key="otau_...") resp = client.chat.completions.create( model="meta-llama/llama-3.3-70b-instruct", messages=[{"role": "user", "content": "Hi"}], ) print(resp.choices[0].message.content)
// npm install opentau import OpenTau from "opentau"; const client = new OpenTau({ apiKey: "otau_..." }); const resp = await client.chat.completions.create({ model: "meta-llama/llama-3.3-70b-instruct", messages: [{ role: "user", content: "Hi" }], }); console.log(resp.choices[0].message.content);

ᛊ  Want it private?

Add -TEE to the model id and the request routes to a sealed hardware enclave — the node operator can't casually read your prompt or its response, and you can verify the exact code that ran via hardware attestation. Same API, one suffix:

model="unsloth/Mistral-Nemo-Instruct-2407-TEE" # sealed
1

Get a key

Sign up and grab an otau_ key at the dashboard. Free tier included.

2

Change the base URL

Point your OpenAI client at api.opentau.ai/v1. Nothing else changes.

3

Ship

Your code now runs on thousands of open models across a decentralized network.

Get your API key → Full docs

OpenTau is OpenAI-API-compatible for chat, models, and streaming. See the docs for the full surface (search, images, code execution, fine-tuning).