PMS Cloud

Build on
PMS Cloud

VMs, containers and an OpenAI-compatible AI API. Docs for people, and for the agents that work with them.

from openai import OpenAI

client = OpenAI(
    base_url="https://api.pmservers.org/ai/v1",
    api_key="pms_sk_...",
)

answer = client.chat.completions.create(
    model="gemma4-e4b",
    messages=[{"role": "user", "content": "Hello"}],
)
print(answer.choices[0].message.content)

import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://api.pmservers.org/ai/v1',
  apiKey: process.env.PMS_API_KEY,
});

const answer = await client.chat.completions.create({
  model: 'gemma4-e4b',
  messages: [{ role: 'user', content: 'Hello' }],
});
console.log(answer.choices[0].message.content);

curl https://api.pmservers.org/ai/v1/chat/completions \
  -H "Authorization: Bearer $PMS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemma4-e4b",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Products

Choose what you build

Start with a VM or the AI API. Containers and storage come next.

VMs

Ubuntu VMs with a browser console. Pick a size, add an SSH key, and start.

AI API

OpenAI-compatible chat, embeddings and decisions. Billed by token.

ContainersComing soon

Always-on workers that connect out. No disk, no inbound traffic.

S3 storageComing soon

S3-compatible buckets. Works with the AWS CLI, boto3 and rclone.

Using an agent?

Give your agent this URL:

https://docs.pmservers.org/llms.txt

It lists every page as a raw Markdown file. All pages in one file are at https://docs.pmservers.org/llms-full.txt. See For agents.

Addresses

What	Address
Console	`https://console.pmservers.org`
AI API (OpenAI-compatible)	`https://api.pmservers.org/ai/v1`
Management API	`https://api.pmservers.org/v1`

Build onPMS Cloud