A2Intelligence.
The AI model from Germany.
A language model with reasoning, image understanding and tool calls. Operated in Germany, available through an OpenAI-compatible API.
- Model
- a2intelligence
- Context window
- 131,072 tokens
- Interface
- OpenAI Responses API
- Operated in
- Germany
What A2Intelligence can do.
Four capabilities, one interface. The examples are interactive.
Reasoning
Thinks in several steps before it answers.
For analysis, planning and code. The reasoning steps can be shown.
We want to automate how our company prepares quotes. Where should we start?
- 01Break down the current process: request → costing → approval → sending
- 02Find the bottleneck: costing is done by hand from price lists
- 03Check the data: item master, price lists and past quotes are available
- 04Weigh the risk: prices and approval must stay with people
- 05Recommendation: start with an AI draft of the costing, the rest later
Start with costing: it is the most time-consuming manual step and relies on data you already have. A2Intelligence drafts it from the request, your price list and past quotes – approval stays with your team.
Image understanding
Understands screenshots, diagrams, photos and documents.
Tool calls
// your application → A2Intelligence { "role": "user", "content": "Book an appointment for Hoffmann GmbH on 2 October at 9:30." }
{ "type": "function_call", "call_id": "call_7f3c", "name": "create_appointment", "arguments": "{\"customer\":\"Hoffmann GmbH\",\"date\":\"2026-10-02T09:30\"}" }
// your code runs the function { "type": "function_call_output", "call_id": "call_7f3c", "output": "{\"status\":\"booked\"}" }
// output_text Done: the appointment with Hoffmann GmbH is booked for Friday, 2 October, 9:30 am.
Calls your functions and APIs in a structured way – the foundation for AI agents.
Large context
Long documents and entire codebases in one request. Pick an example:
- Used (estimate)
- ≈ 92,000
- Left for the answer
- ≈ 39,072
- Utilisation
- 70 %
Estimates for illustration – the actual token count depends on language and content.
Processing
in Germany.
Requests to A2Intelligence are processed in Germany. The platform servers are located in Germany as well.
Located in Germany
The model is operated in Germany, not at US hyperscalers.
No training on your data
Your input is only used to answer your request.
What companies
use it for.
Through the API in your software – or as a solution we build with you.
Software development
Coding agents, code reviews and documentation.
With a2agent, our coding agent built on A2Intelligence.
Assistants
Internal knowledge assistants and customer service.
Answers from your own documents – with sources.
Automation
Analyse documents, extract data, orchestrate processes.
From the inbox into your system – with human approval.
Integration by our team
We connect A2Intelligence to your existing systems.
Try
A2Intelligence.
Create an account, create an API key, send a request. Or try it without code in the platform playground.
Read more
OpenAI-compatible API for A2Intelligence
Connect A2Intelligence through the OpenAI Responses API: examples for curl, JavaScript and Python, with Chat Completions still supported as a legacy endpoint.
AI developer platform by A2Genesis
The A2Genesis platform is self-service access to A2Intelligence: create API keys, top up credit or subscribe, and track usage in your dashboard.
LLM API pricing and plans
Transparent pricing for A2Intelligence: pay-as-you-go credit with no base fee, or the monthly Starter and Pro subscriptions with generous usage limits.
Why AI made in Germany matters for businesses
Privacy, independence and transparency: why businesses should use AI models running on servers in Germany and what to look out for when choosing one.