Menu
Get started

Getting started

Create an account, add a payment method and deploy your first database or model.

This guide takes you from a new account to a running database or model on the On-Demand plan.

Create an account#

  1. Go to nistro.ai/signup.
  2. Enter your name, email, company and a password of at least 8 characters, then select Sign up. Signing up accepts the Terms of Service and Privacy Policy.
  3. Open the verification email and follow its link to activate your account.
  4. Sign in with your email and password.

You can also sign up and sign in with Google or GitHub.

Add a payment method#

You need a payment method on file before you can deploy. Until you add one, the console shows a banner, and a deploy fails with "A payment method is required before deploying. Add one under Billing."

  1. In the console, open Billing.
  2. On the Payment Method card, select Configure. This opens the billing provider's page, where you add your payment details.

When the card shows Configured, you're ready to deploy.

Deploy a database#

  1. Open Postgres Deployments and select Create.
  2. Keep the On-Demand plan. Set vCPU and storage; memory follows vCPU at 4 GiB per vCPU.
  3. Optionally pick a workload preset, such as OLTP, and change the database name and username (both default to app). Leave the password blank to have one generated.
  4. Keep Location on Any to let Nistro AI pick the lowest-priced location, or choose US, EU or Asia. Then select Load pricing.
  5. Review the offer and select Deploy.
The On-Demand requirements step: capacity sliders, workload presets and database settings

The deployment's page opens. When its status is Running, the Quick test card has a ready-made psql command:

Shell
psql 'postgresql://<username>:<password>@<host>:5432/<database>'

See Nistro AI Postgres for connection details and everything else you can configure.

Deploy a model#

  1. Open Inference Deployments and select Create.
  2. Keep the On-Demand plan. Enter a Hugging Face model id, such as google/gemma-4-26B-A4B, or paste its huggingface.co URL. The console checks the model as you type.
  3. Keep Location on Any, or choose US, EU or Asia, and select Recommend hardware.
  4. Review the recommended hardware, select View pricing, then select Deploy.
The Hardware step with the required VRAM, topology, model format and sizing details

When the deployment is ready, its Overview shows the OpenAI-compatible endpoint, and the Call the API tab shows the API key and example requests. Save the key as NISTRO_TOKEN, then send a request:

Shell
curl <endpoint>/openai/v1/chat/completions \
  -H "Authorization: Bearer $NISTRO_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"model":"<model>","messages":[{"role":"user","content":"Hello"}]}'

Replace <endpoint> with the deployment's endpoint and <model> with the model name shown in the console's examples. See Nistro AI Inference for the OpenAI SDK and streaming.

Next steps#