For developers
One key.
Every model.
Keep the SDK you already use. Change the base URL and the model name, and you are on HFAI Cloud.
Drop-in compatibility.
The same request and response shape
your clients already speak.
Keep the SDK you already use. Change the base URL and the model name.
HFAI Cloud serves POST /v1/chat/completions and GET /v1/models in the OpenAI chat-completions shape, streaming included. Every OpenAI-compatible SDK, client and harness works by changing two strings: the base URL, https://api.heartflowai.net/v1, and the key, hfai_live_….
Switching models is one edit inside the request body — the model string. There is no per-model SDK, no second account and no second balance to top up when you move on.
curl, Python
and JavaScript
Three complete calls againsthttps://api.heartflowai.net/v1.
CURL
curl https://api.heartflowai.net/v1/chat/completions \
-H "Authorization: Bearer hfai_live_…" \
-H "Content-Type: application/json" \
-d '{"model":"deepseek-v4-pro","messages":[{"role":"user","content":"Hello HFAI"}]}'
PYTHON
# pip install openai
from openai import OpenAI
client = OpenAI(
base_url="https://api.heartflowai.net/v1",
api_key="hfai_live_…",
)
response = client.chat.completions.create(
model="deepseek-v4-pro",
messages=[{"role": "user", "content": "Hello HFAI"}],
)
print(response.choices[0].message.content)
JAVASCRIPT
// npm install openai
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.heartflowai.net/v1",
apiKey: "hfai_live_…",
});
const response = await client.chat.completions.create({
model: "deepseek-v4-pro",
messages: [{ role: "user", content: "Hello HFAI" }],
});
console.log(response.choices[0].message.content);
Replace hfai_live_… with a key from HFAI Cloud and deepseek-v4-pro with any model id from https://api.heartflowai.net/v1/models.
Use it from your coding agent.
An OpenAI-compatible endpoint
for the tools you already run.
Agent tools and editors that accept a custom OpenAI-compatible provider take the same two values: the base URL https://api.heartflowai.net/v1 and an hfai_live_… key. Add them as a new provider alongside the built-in ones, choose a model from the catalogue, and the agent runs on your HFAI Cloud balance. Nothing else in the tool’s configuration changes, and no provider-specific integration is needed per model.
What you get
One endpoint for every model
in the catalogue.
- One balance. Credits bought once with
$HFAIat a USD peg and drawn down as you call models. - One key. A single
hfai_live_…key works for every model, in every client. - No markup on inference. You pay the model’s list price — no tiers, no discounts, nothing added on top when you buy credits.
- Prepaid credits. No subscription, no per-seat licence, no monthly minimum.
- The same endpoint for every model. Switching models is changing one string in the request body.
Change two strings.
Keep your stack.
Sign in with a Solana wallet, fund your balance with $HFAI, and mint your API key.