Issue 01 — a VPS for language models · sign-up open
Not the tokens.
A private LLM instance on EU hardware, for a flat monthly fee. Nothing counts down and nothing resets — the only limits are how many requests you run at the same moment and how much context each one gets, and you set both yourself.
No token meter.
No rolling usage window.
No request cut off halfway
through your task.
No surprise bills — your instance
at one fixed monthly cost.
EU hardware, EU jurisdiction.
01 / The idea
If you've bought a server, you know this product. #
One project gets one VPL: one endpoint, one key, two numbers to size it.
A slider sets how many instances the project runs, and an instance is one request in flight. Three instances means three things at once: your agent, your editor, your test loop. A second dial sets the context window each of those requests gets. The API is OpenAI-compatible, so Cline, ZooCode, VS Code BYOK and your own backend all talk to it unmodified.
01 On a VPS — vCPU + RAM
Instance count
One slider, one dial. Inside that size, run as much as you like, all day.
02 On a VPS — no request quota
No usage window
Nothing counts down and nothing resets. You are never locked out mid-task.
03 On a VPS — flat monthly invoice
Flat monthly invoice
No surprise bills: a busy sprint costs exactly what a quiet one costs. Tokens are counted for you and never billed to you.
04 On a VPS — OS image
Open-weight model
Pick and pin the model. Swap it by changing one string, no migration.
05 On a VPS — one server, one host
One project, one VPL
Keys, usage and limits stay per project — nothing is shared by accident.
06 On a VPS — datacenter region
EU-only region
Nothing leaves EU jurisdiction — GDPR and AI Act posture by construction.
02 / Migration
Two minutes, two fields. #
Nothing to install, nothing to rewrite.
01
Create a project
A project is one VPL: its own endpoint, its own key, its own usage. Name it after the thing it serves.
02
Set the instance count
One slider decides how many requests run at once. Move it up for a week of parallel agents, back down after. That is the whole capacity model.
03
Point your tool at it
Base URL and key into Cline, ZooCode, VS Code BYOK, or any OpenAI SDK. Nothing else in your setup changes.
export OPENAI_BASE_URL=https://api.solheim.ai/v1
export OPENAI_API_KEY=slh_live_9f3c…
03 / Sovereignty
EU-hosted, all the way down. #
Sovereignty is an architecture, not a checkbox.
01 Infrastructure
GPUs at EU-owned providers in EU datacenters. No US hyperscaler in the serving path, so no CLOUD Act exposure to write into your DPA.
02 Models
Open weights under permissive licenses, pinned per instance and run by us. Your prompts are not training data — not ours, not anyone's.
03 Routing
Requests never leave the region, and this page loads nothing from a third-party origin — no US CDN, no analytics, no fonts phoning home.
The library
01 Qwen3.6-35B-A3B
qwen3.6-35b-a3b
The default for day-to-day work: diffs, tests, refactors and tool-calling agents. A mixture-of-experts model: 35B parameters in total, three billion of them active per token, price-value like nothing else.
02 DeepSeek V4 Flash
deepseek-v4-flash-0731
For more demanding work: near frontier open-weight performance on your own reserved compute
Request a model
open weights · permissive licence
04 / Pricing
Two dials: how many, how much context. #
Instances × context window × a flat monthly rate. Still no tokens, no tiers and no overage — there is nothing on this invoice that can surprise you.
- Your window
- 64k
- Model
- one per project
- Beyond your count
- queues, never 429
- Overage
- there isn't any
€15.00 /month · 1 × 64k × €15.00
Excludes VAT, which is added at checkout based on where you are.
Start at this size
Beyond ten, or beyond 128k
Dedicated & enterprise
A dedicated GPU or a MIG slice of one — no shared neighbours, no contention with anyone else's agents. A 256k window lives here too: it is four times the cache of the largest shared size, so it gets its own card rather than a queue. Private networking, a signed DPA and invoicing rather than a card. Priced against the hardware you need, not per request.
Let's talk
05 / Readership
Who it's for. #
Same product, same dials, same arithmetic.
01 Developers running coding agents
You live in Cline, ZooCode or VS Code and you hit the window most weeks. Three instances is an agent, an editor and a test loop running at once — for as long as the day lasts.
02 Founders who need an inference backend
Your product needs a private LLM behind it, with a bill that doesn't move when traffic does. Point your backend at the same endpoint and size it the same way.
It is one product, not two. What changes between these is what you point at the endpoint — not the pricing, not the API, not the machine.
06 / From the blog