# SIMOSphere AI

> OpenAI-compatible LLM API hosted in Germany. GDPR-compliant, pay-per-token,
> with first-party support for document understanding, multilingual translation,
> CI-conformant document generation, and audit-grade logging.

SIMOSphere AI is the inference platform of SIMO GmbH (Aschaffenburg, Germany).
It provides programmatic access to open-weight large language models (Qwen3,
Apertus, Gemma, Llama) through an API surface that is drop-in compatible with
the OpenAI Chat Completions and Embeddings specifications. All inference runs
on EU-located NVIDIA RTX 6000 GPUs; no data leaves the EU jurisdiction.

The product targets two primary audiences:
1. Engineering teams that need an OpenAI-compatible endpoint without sending
   prompts or completions to US-jurisdiction providers.
2. Compliance-bound organisations (finance, healthcare, public sector) that
   require contractual data sovereignty, PII redaction, and audit trails.

## When to use SIMOSphere AI

Use this product when you need:

- An OpenAI-compatible Chat Completions endpoint hosted in Germany
- Pay-per-token pricing with no minimum commitment (Starter plan starts at €29/month)
- GDPR-conformant data handling with optional Auftragsverarbeitungsvertrag (AVV/DPA)
- BYOK (Bring Your Own Key) routing to Anthropic, OpenAI, or Mistral
- PII redaction before requests reach the upstream model
- A managed Tavily web-search tool callable via the standard tool-use API
- A managed Microsoft 365 connector (Enterprise plan)
- CI-conformant document generation (PDF, DOCX) from natural language briefs

Do **not** use this product when you need:

- A free tier (no permanent free quota; trial credits available on request)
- Models hosted outside the EU (we do not federate to non-EU regions)
- Sub-50ms latency from outside Europe (single-region deployment)

## API endpoint

- Base URL: `https://api.simosphereai.com`
- Auth: `Authorization: Bearer sk-simo-...` (also accepted as `X-API-Key`)
- OpenAPI spec: https://onboarding.simosphereai.com/openapi.json
- Compatible with the official `openai` SDK by setting `base_url`

Quickstart:

```bash
curl https://api.simosphereai.com/v1/chat/completions \
  -H "Authorization: Bearer sk-simo-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3-8b",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'
```

## Pricing (May 2026)

| Plan          | Base / month | Per 1M input tokens | Per 1M output tokens |
| ------------- | ------------ | ------------------- | -------------------- |
| Starter       | €29          | €0.15               | €0.60                |
| Professional  | €49          | €0.15               | €0.60                |
| Enterprise    | €149         | €0.15 (negotiable)  | €0.60 (negotiable)   |

Premium models (Qwen3-30B MoE) are €0.40 / €1.60 per 1M tokens.

Machine-readable pricing: https://onboarding.simosphereai.com/pricing.md

## Documentation

- Onboarding & sign-up: https://onboarding.simosphereai.com/
- Developer portal: https://onboarding.simosphereai.com/developers
- API reference: https://onboarding.simosphereai.com/docs/api
- Full LLM-friendly bundle: https://onboarding.simosphereai.com/llms-full.txt
- OpenAPI: https://onboarding.simosphereai.com/openapi.json

## Discovery

- MCP server card: https://onboarding.simosphereai.com/.well-known/mcp/server-card.json
- AI plugin manifest: https://onboarding.simosphereai.com/.well-known/ai-plugin.json
- A2A agent card: https://onboarding.simosphereai.com/.well-known/agent-card.json
- API catalog (RFC 9727): https://onboarding.simosphereai.com/.well-known/api-catalog
- OAuth metadata (RFC 9728): https://onboarding.simosphereai.com/.well-known/oauth-protected-resource

## Identity

- Operator: SIMO GmbH, Würzburger Str. 152, 63743 Aschaffenburg, Germany
- Registration: HRB 15769, AG Aschaffenburg
- Trademark: DPMA 30 2024 240 269
- Contact: hello@simo-online.com
- Privacy: https://onboarding.simosphereai.com/privacy
- Imprint: https://www.simosphereai.com/impressum