canonical: https://jentic.com/apis/deutschlandgpt.de/deutschlandgpt

# Deutschlandgpt De DeutschlandGPT API

Jentic publishes the only available OpenAPI specification for DeutschlandGPT API, keeping it validated and agent-ready. DeutschlandGPT is a GDPR-compliant AI inference provider hosted in Germany, exposing OpenAI-compatible endpoints for chat completions, embeddings, image generation, and model listing. Drop-in compatibility means existing OpenAI client code can be repointed at DeutschlandGPT for workloads that require EU data residency or stricter privacy posture.

## For AI agents

Run chat completions, generate embeddings, create images, and list models on DeutschlandGPT - a GDPR-compliant, OpenAI-compatible inference endpoint hosted in Germany.

## Scope

Does not handle audio transcription, fine-tuning jobs, or assistants beyond chat, embeddings, and images - use for OpenAI-compatible inference only.

## Capabilities

- Run chat completions against German or multilingual models with OpenAI-compatible payloads
- Generate vector embeddings for semantic search and retrieval workflows under GDPR
- Create images from text prompts with the OpenAI-style images endpoint
- List the catalogue of available models hosted on DeutschlandGPT
- Repoint existing OpenAI SDK clients at DeutschlandGPT by changing the base URL

## Use cases

### GDPR-Compliant Chat Completions

Companies with EU customers or strict privacy requirements use DeutschlandGPT as a drop-in replacement for OpenAI to keep inference traffic on German infrastructure. The `/v1/chat/completions` endpoint accepts the standard OpenAI request shape, so application code rarely needs more than a base URL change to switch over.

Example prompt: POST `/v1/chat/completions` with a German-language system prompt and the user's question, then return the assistant message.

### RAG with EU Data Residency

Retrieval-augmented generation pipelines that index sensitive customer documents use the `/v1/embeddings` endpoint to keep both indexing and retrieval inside GDPR-compliant infrastructure. Pair embeddings with a vector database hosted in the EU to keep the whole loop on-region.

Example prompt: POST `/v1/embeddings` with a batch of 50 document chunks and store the returned vectors in an EU-hosted vector database.

### Image Generation for German Marketing

Marketing teams use `/v1/images/generations` to produce visuals for German-language campaigns without sending prompts or assets to non-EU infrastructure. Useful for regulated industries - banks, healthcare, public sector - where data movement to the US is a compliance issue.

Example prompt: POST `/v1/images/generations` with a German prompt describing the campaign visual and download the resulting image to the asset library.

### Agent-Driven EU AI Workflows

AI agents serving European users can route inference through DeutschlandGPT via Jentic to satisfy data-residency and GDPR posture requirements. Jentic stores the bearer token in the vault and matches agent intents to the right OpenAI-compatible endpoint.

Example prompt: Search Jentic for 'run a chat completion on a GDPR-compliant model', load the `/v1/chat/completions` operation, and execute it with the user's prompt.

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| POST | `/v1/chat/completions` | Run a chat completion |
| POST | `/v1/embeddings` | Generate vector embeddings |
| POST | `/v1/images/generations` | Generate an image from a prompt |
| GET | `/v1/models` | List available models |

## Key resources

- **Chat Completions** — OpenAI-compatible chat completion endpoint for German and multilingual models
- **Embeddings** — Generate vector embeddings for semantic search
- **Images** — Generate images from text prompts
- **Models** — List available models hosted on DeutschlandGPT

## Why Jentic

- **Setup:** Wiring DeutschlandGPT by hand means learning its bearer auth and building the OpenAI-compatible chat, embeddings, image-generation, and models calls yourself. Through Jentic you install once, import the DeutschlandGPT API from the API Directory, store the token once, and your agent calls it.
- **Permission scoping:** DeutschlandGPT takes the prompt and parameters in the request body, so you limit the agent to the operations it needs, such as chat completions or embeddings, and image generation is included only if you add that operation.
- **Credential handling:** Your DeutschlandGPT bearer token is stored once, encrypted, by your own Jentic One instance and injected into the Authorization header at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'run a GDPR-compliant chat completion' or 'embed this text', and Jentic returns the matching DeutschlandGPT operation with its input schema so the agent calls the right endpoint without reading the docs.

## Related APIs

- **OpenAI API** — OpenAI is the request-shape reference; DeutschlandGPT mirrors it on EU infrastructure.
- **Anthropic Messages API** — Anthropic Messages is a non-OpenAI-shape alternative with different model strengths.
- **Replicate API** — Use Replicate for hosted open-source models alongside DeutschlandGPT for primary inference.

## FAQ

### Why is there no official OpenAPI spec for DeutschlandGPT API?

DeutschlandGPT does not publish an OpenAPI specification. Jentic generates and maintains this spec so that AI agents and developers can call DeutschlandGPT API via structured tooling. It is validated against the live API and kept up to date. Get started with Jentic One, the self-hosted execution layer.

### What authentication does the DeutschlandGPT API use?

DeutschlandGPT uses bearer token authentication, identical to the OpenAI scheme. Through Jentic the token is stored encrypted in the vault and injected into the Authorization header at call time.

### Is the DeutschlandGPT API OpenAI-compatible?

Yes - the chat completions, embeddings, images, and models endpoints follow the OpenAI request and response shapes, so existing OpenAI SDK clients usually only need a base URL change to talk to DeutschlandGPT.

### Can I generate vector embeddings for RAG with the DeutschlandGPT API?

Yes - POST `/v1/embeddings` returns vectors that can be stored in any vector database. Combined with EU-hosted vector storage, the entire retrieval-augmented generation loop stays within GDPR-compliant infrastructure.

### What are the rate limits for the DeutschlandGPT API?

The OpenAPI spec does not declare specific rate limits. Treat them as comparable to other hosted LLM APIs and implement exponential backoff on 429 responses.

### How do I run a chat completion through Jentic?

Search Jentic for 'run a chat completion on DeutschlandGPT', load the `/v1/chat/completions` operation, and execute it with the messages array. Jentic injects your bearer token at execution time.

### Can I limit what my agent is allowed to do with the DeutschlandGPT API?

Yes. Because you run Jentic One yourself, your own rules decide which DeutschlandGPT operations the agent can call, so you can allow only chat completions and embeddings while leaving out image generation and the models list. Image generation is available to the agent only if you explicitly add the `/v1/images/generations` operation. Your bearer token is held by your own Jentic One instance and injected into the Authorization header at call time, so the agent never sees the credential and can only reach the operations you have granted.