LLMBase
OpenAI-compatible inference API · EU-hosted

The inference API for Europe.

Leading open models through an OpenAI-compatible API: EU-hosted, with a DPA and billed per token. Swap the base URL, add your API key, done.

Create API keyRead the docs
  • from $0.18 / 1M input tokens
  • Prepaid from $5
  • DPA included

Code example: the OpenAI Python SDK pointed at https://api.llmbase.ai/v1.

Trusted by companies of all sizes and industries

  • Operated by a German company
  • EU data centers
  • DPA under Art. 28 GDPR
  • No training on your content

One provider, three paths

Choose how you use AI.

Start with the API, move to your own dedicated endpoint when you need it, and give your team the chat on top.

For developers and product teams

Inference API

Build EU-hosted open models into your product, scripts or agents with the OpenAI SDK. You only pay for what you use.

  • OpenAI SDK compatible
  • Streaming and tool calls
  • Usage-based billing
  • Model catalog API

from $0.18 / 1M input tokens

Prepaid credit from $5, no minimum term

Create API keyModels and pricing

Hosted in Germany

Dedicated endpoints

Dedicated capacity including operations. Your own endpoint with the same API, operated by us, with SLA and monitoring.

from €2,690 / month net

Subject to capacity confirmation.

View packages

For your whole team

AI Chat

A private AI chat with web search, files and shared workspaces, without writing a line of code.

Start free, Pro €19 / month

Open chatCompare plans

Inference API

Three steps to your first request.

No sales call, no waiting list. You keep your code and only swap the base URL.

  1. 1

    Create an account and an API key.

    The key works for every model available over the API as soon as you have credit.

  2. 2

    Top up credit.

    From $5, manually or with automatic top-up. API credit is separate from the chat subscription.

  3. 3

    Swap the base URL.

    https://api.llmbase.ai/v1
Create API keyRead the quickstart

Available over the API today

EU-hosted
All models, prices and limits

Key management API

API keys for every customer, created by your own systems.

Your ERP or backend creates, limits and revokes Inference API keys through /v1/keys, for example one key per end customer, all paid from your prepaid credit.

  • Up to 5 active management keys per account.
  • Management keys cannot run inference; inference keys cannot call /v1/keys (both 403).
Create management keyKey management docs
POST /v1/keysCreate a key with a monthly limit
curl https://api.llmbase.ai/v1/keys \
  -H "Authorization: Bearer $LLMBASE_MANAGEMENT_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Customer 4711",
    "monthly_limit_usd_cents": 50000,
    "daily_limit_usd_cents": null
  }'

50000 cents = USD 500 per UTC month. Store the management key like a password; the example reads it from an environment variable.

  • Create keys via API

    POST /v1/keys creates an Inference API key and returns its secret once. Authenticate with a management key (llmbase_mk_…) that you create in the dashboard.

  • Monthly and daily limits

    Optional per key: monthly_limit_usd_cents and daily_limit_usd_cents, a whole number of USD cents from 1 to 1,000,000 (USD 10,000), or null for no limit.

  • Audit event on every change

    Every create, change and revoke writes an audit event in the same database batch as the change itself.

Screenshot of the API keys page in the LLMBase dashboard with demo data: inference keys with daily and monthly credit caps, and the management keys section with its security notice, a create form and two demo management keys, ERP and Billing, shown only by their masked prefix.
Demo data: management keys and key limits in the dashboard, prefixes masked.

Managed & Dedicated Endpoints

API with location Germany.

Need your requests processed in Germany only? On request, the API with location Germany processes your Inference API requests exclusively in Germany.

  • The same OpenAI-compatible API, your code stays unchanged
  • Processing in Germany only
  • GDPR, DPA and technical and organisational measures (TOMs)
  • Requested through the enterprise form, with an offer for your project
Request API with location GermanyView dedicated endpoints

How the price is made up

Inference API, EU-hosted
Normal price
API with location Germany
+ 20 %
All other prices
unchanged

Available on request. All other prices stay unchanged.

From Germany. Trusted across Europe.

LLMBase is operated by a German company, and personal data is processed under the GDPR and German data protection law. Teams across the EU get a DPA with technical and organisational measures, EU-hosted open models, and hosting details for every model.

  • GDPR
  • DPA available
  • No training on your data
Read our privacy policy
LLMBase · GermanyEU member states

Built for the work your teams do.

From the shop floor to the clinic: use the chat for everyday tasks and the API to build AI into your own processes.

  • Humanoid robot operating a CNC machine on a factory floor

    Robotics

    Turn natural-language instructions into structured robot commands via the API, and let engineers query logs and specs in chat.

  • Healthcare & clinical documentation

    Draft discharge letters and visit notes from dictation, and summarise long case histories for the next shift.

  • Legal

    Review contracts for unusual clauses, compare versions, and prepare first drafts of letters and memos.

  • Manufacturing

    Search maintenance manuals in chat, and classify shift reports and quality notes automatically through the API.

  • Customer service

    Suggest replies from your help-centre content, route incoming tickets, and summarise long conversations for hand-off.

  • Software development

    Explain unfamiliar code, write tests, and plug models into your own tools with the OpenAI-compatible API.

  • Education

    Prepare lesson material and exercises, give structured feedback on drafts, and explain difficult topics step by step.

  • Your use case is not listed? Build it with the API or try it in the chat.

    Create API keyStart chatting

For business customers

Your DPA, ready to sign.

Business customers can conclude a Data Processing Agreement (DPA) with Eyloo GmbH, the operator of LLMBase.

  • Concluded with Eyloo GmbH, the operator of LLMBase
  • A contract for commissioned processing under Art. 28 GDPR
  • Available directly in your account
Sign your DPATalk to salesView pricing

News

What’s new at LLMBase

Product releases, research, and company updates.

All news
product

LLMBase goes open-weight

At LLMBase, we are strengthening privacy, security and transparency.

Read update
product

Introducing Computer: a bounded workspace for complex AI tasks

LLMBase Chat can now inspect authorised files, unpack public ZIP archives, review public code, and turn multi-step work into a finished deliverable.

Read update
product

LLMBase Chat now supports Microsoft Office files

Bring Word documents, Excel workbooks, and PowerPoint presentations into LLMBase Chat to understand, compare, and move work forward.

Read update
product

LLMBase Chat can now create and edit files in one workspace

Move from a conversation to a reusable report, spreadsheet, presentation, or structured file—and keep the surrounding project context organised.

Read update

Common questions before you start.

What does the Inference API cost?

You pay per token, from $0.18 / 1M input tokens, depending on the model; output tokens cost more. The API runs on prepaid credit from $5. There is no base fee and no subscription.

Where are my requests processed?

Every Inference API model is hosted by LLMBase in EU data centers. LLMBase is operated by a German company, and a DPA is available for business customers.

Do I need a chat subscription for the API?

No. API credit and the chat subscription are two separate budgets. API requests only use your prepaid credit, and chat usage does not reduce your API credit.

When do we talk in person?

For projects from €10,000 per year or for the API with location Germany. Describe your project in the enterprise form and we get back to you with an offer. Below that, start right away with an API key.

Enterprise form

Your first request, EU-hosted.

Run by a German company, with a DPA and hosting details per model. Create an API key and get started.

Create API keyRead the docs

A project from €10,000 per year or for the API with location Germany? Request a project