Get Started
InertialAI is two products behind one API key:
- InertialAI-0.1 — the chat model for real-world data (sensor streams, metrics, markets, health signals), behind an OpenAI-compatible API.
- Chronicle — self-serve finetuning: train on your data from $5, with an automatic before/after eval and a private scale-to-zero endpoint.
This page gets you from "I have an email address" to "the model just read my sensor data".
1. Create an account
- Visit app.inertialai.com and click Sign up.
- Fill out the registration form (email, name, password) and accept the terms.
- Enter the verification code sent to your email and click Verify email (check spam if it doesn't arrive).
You'll land in the dashboard, signed in. Every new account starts with $5 of free usage — enough to chat with the model at length, embed, forecast, and train a first Chronicle finetune.
2. Create an API key
The API authenticates with API keys (they look like iai_...) sent as an
HTTP bearer token.
- In the dashboard, open the API Keys tab.
- Click Create API Key, give it a title, and click Create Key.
- Copy the key immediately — it is shown only once.
Set it as an environment variable so clients pick it up automatically:
export INERTIALAI_API_KEY="iai_..."
Every request then carries it in the Authorization header:
Authorization: Bearer iai_...
- API base:
https://inertialai.com/api/v1 - Payloads: JSON request/response bodies
3. Make your first request
InertialAI-0.1 is OpenAI-compatible, so the standard OpenAI Python client is all you need:
pip install openai
import os
from openai import OpenAI
client = OpenAI(
base_url="https://inertialai.com/api/v1",
api_key=os.environ["INERTIALAI_API_KEY"],
)
resp = client.chat.completions.create(
model="inertialai-0.1",
messages=[
{
"role": "user",
"content": (
"Hourly server CPU temps (C): 61, 62, 61, 63, 88, 91, 90. "
"What happened, and should I worry?"
),
}
],
)
print(resp.choices[0].message.content)
print(resp.usage) # exact token counts = your exact cost basis
:::tip Cold starts The model scales to zero when idle, so the first request after an idle period can take up to ~5 minutes (measured). Send a 1-token warmup ping first if that matters to you — see the quickstart. :::
The same request as raw HTTP, if you want to see the wire shape:
curl -sS \
-X POST "https://inertialai.com/api/v1/chat/completions" \
-H "Authorization: Bearer ${INERTIALAI_API_KEY}" \
-H "Content-Type: application/json" \
-d '{
"model": "inertialai-0.1",
"messages": [
{"role": "user", "content": "Readings: 61, 62, 88, 91. Normal or anomaly?"}
]
}'
Troubleshooting the first call
| Symptom | Meaning | Fix |
|---|---|---|
| First request hangs for minutes | Cold start — the GPU is spinning up and the engine is compiling | Expected after idle; keep client timeouts ≥ 360 s, or send a warmup ping |
401 Unauthorized | Key missing or malformed | Header must be exactly Authorization: Bearer iai_... |
403 Forbidden | Key invalid, revoked, or expired | Create a fresh key in API Keys |
422 Validation Error | Body doesn't match the schema | Include model: "inertialai-0.1" and a messages array; compare against the Chat Completions reference |
429 Too Many Requests | Sending too fast | Back off and retry; honor Retry-After |
Stuck? support@inertialai.com.
Where to go next
- InertialAI-0.1 in depth — request/response contract, streaming, token-level confidence you can gate on: Overview · Chat Completions API · Token-Level Confidence
- Chronicle — finetune on your data and deploy privately: Chronicle overview · Finetuning Quickstart
- Billing — credits, per-token rates, GPU serving rates: Billing & credits
- Other endpoints — embeddings, forecasting, and the Python SDK: Embeddings · Forecasting · Python SDK