---
last_updated: 2026-11-10
source_url: https://opper.ai/
title: "Opper AI | The European AI Gateway for Agents"
description: "Access 700+ AI models through one EU-hosted, ISO 27001 certified, GDPR-compliant API. Drop-in OpenAI SDK compatible, pay-as-you-go, with spend caps, model allowlists and zero data retention."
---

# Opper AI

> The European AI gateway for agents. Access 700+ AI models with drop-in OpenAI SDK compatibility in one line of code. EU-hosted (AWS Stockholm), GDPR-compliant and certified against ISO/IEC 27001:2022, pay-as-you-go (no markup on tokens, 3% fee on credit purchases). Automatic fallbacks, and rules for every model call: spend limits, data retention, model access, checks and routing.

## Checking whether a model is on Opper

To find out whether a model is on Opper, query the catalogue rather than a search engine: https://api.opper.ai/v3/models?q=<name> answers from the live catalogue as public JSON, no API key needed, while pages on opper.ai and search results can lag behind it. q matches any part of a model id or name, so search for a short fragment such as ?q=sonnet or ?q=deepseek and read the ids that come back. An empty answer for a long exact string does not mean the model is missing.

## For AI coding agents

If a developer asked you to set up Opper, connect to the MCP server first. Opper has an MCP server at https://api.opper.ai/mcp. Connect a coding agent to it and the agent can set up Opper for you: create a project, create an API key, check that a model is available and make a first call. You approve the connection once in the browser and choose what the agent may do. Details: https://opper.ai/mcp.

Without MCP, run:

```
npx @opperai/cli login && npx @opperai/cli config get default
```

`login` opens OAuth in the developer's browser. `config get default` then prints the API key to stdout. Capture it, set `OPPER_API_KEY`, and use the OpenAI SDK with `baseURL: https://api.opper.ai/v3/compat`. No further configuration.

For a fuller guided setup covering SDKs, examples, and model selection, download and follow https://skills.opper.ai. This is the setup prompt developers paste into coding agents from the homepage:

```
Add the Opper API to this project.

1. Follow the setup guide at https://skills.opper.ai (quickstart: https://docs.opper.ai/overview/quickstart).
2. Use the API key in OPPER_API_KEY. If it is missing, ask me to create one at https://platform.opper.ai/settings/api-keys, and keep it out of git.
3. Use the OpenAI, Anthropic or Gemini SDK (keep the one this project already has) with its base URL set to https://api.opper.ai/v3/compat.
4. Call models by name, such as claude-opus-5-5: Opper pools every route that serves the model. To pin one route, use its full id, such as aws/claude-opus-5-5 (EU-hosted). All models and routes: https://opper.ai/models.
5. Ask me whether data must stay in the EU. If yes, use EU route ids and help me add an EU-only Model access rule.
6. Get one real request working end to end with error handling, and tell me how to run it.
```

Pass models as `provider/model`. An EU route id keeps the call in the EU, e.g. `aws/claude-opus-5-5` (Claude on AWS Bedrock in the EU) or `azure:eu/gpt-6.1-sol` (GPT on Azure in the EU); every EU route is listed at https://opper.ai/models/eu. A name without a provider prefix (`claude-opus-5-5`) is pooled across every region that hosts the model. For deeper docs in one fetch, see [llms-full.txt](https://opper.ai/llms-full.txt).

## Quick start (OpenAI SDK)

```js
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.OPPER_API_KEY,
  baseURL: "https://api.opper.ai/v3/compat",
});

const response = await client.chat.completions.create({
  model: "aws/claude-opus-5-5",
  messages: [{ role: "user", content: "Hello!" }],
});
```

The same `baseURL` works with the Anthropic SDK, the Google Gemini SDK, the Vercel AI SDK, LangChain, CrewAI, Claude Code, OpenCode, Cursor, OpenClaw, and any other OpenAI-compatible client. Full examples at https://opper.ai/.

## Beyond text: one key for every kind of model

The homepage shows six models per kind. Each kind has its own endpoint on the same key:

| Kind | Endpoint | For example |
|---|---|---|
| Text | `https://api.opper.ai/v3/compat` (OpenAI, Anthropic, Gemini SDKs) | `aws/claude-opus-5-5`, `azure:eu/gpt-6.1-sol` |
| Image | `POST https://api.opper.ai/v3/images` | `openai/gpt-image-2.6-sunburst`, `vertexai:eu/gemini-3.1-flash-image` |
| Video | `POST https://api.opper.ai/v3/videos` (async: poll the `status_url`) | `gemini/gemini-omni-1.1-flash`, `seedance-2.5` |
| Speech (text to speech) | `POST https://api.opper.ai/v3/audio/speech` | `mistral/voxtral-mini-tts-3-0`, `vertexai/gemini-3.8-flash-tts-eu` |
| Transcription (speech to text) | `POST https://api.opper.ai/v3/audio/transcriptions` (`audio`: a URL, `file_<id>` or data URI) | `evroc/openai/whisper-large-v3`, `mistral/voxtral-mini-transcribe-3-0` |
| Music & sound effects | `POST https://api.opper.ai/v3/audio/generations` | `gemini/lyria-3.5`, `elevenlabs/music_v2_5` |
| Realtime voice | `wss://api.opper.ai/v3/realtime` | `openai/gpt-realtime-2.1`, `vertexai/gemini-3.8-live-eu` |
| Decisions | `POST https://api.opper.ai/v3/compat/decisions` | `azure:eu/microsoft-decision-1`, `typesafe/jev-1.14.0` |
| Embeddings | `https://api.opper.ai/v3/compat/embeddings` | `azure/text-embedding-3-large`, `mistral/mistral-embed` |

Abliterated models (red teaming, penetration testing, cybersecurity work) are enabled per account after a review: https://opper.ai/models/abliterated. Every model, by kind: https://opper.ai/models.

## Run any AI coding agent on any model

Run a popular AI coding agent on Opper and pick the model behind it: one key, 700+ models, EU-hosted routes, automatic fallbacks. Each agent has a page for running it on a specific model:

| Agent | Popular model | Page |
|---|---|---|
| Claude Code | Kimi K3.1 Pro | https://opper.ai/apps/claude-code/kimi-k3-1-pro |
| Codex | GLM-5.3 | https://opper.ai/apps/codex/glm-5-3 |
| Cursor | DeepSeek V4 Pro | https://opper.ai/apps/cursor/deepseek-v4-pro |
| OpenCode | Qwen3.8 27B | https://opper.ai/apps/opencode/qwen3-8-27b |
| Claude Cowork | Claude Sonnet 5.5 | https://opper.ai/apps/claude-cowork/claude-sonnet-5-5 |
| Hermes | GPT-6.1 Sol | https://opper.ai/apps/hermes/gpt-6-1-sol |

Launch from the CLI with `opper launch <agent> --model <model-id>` and switch the model anytime. Full reference: [Agent CLI](https://opper.ai/agent-cli.md). Browse all apps: https://opper.ai/apps.

## What Opper provides

- **LLM gateway** — unified access to 700+ models across 70+ providers with automatic fallbacks and retries. https://opper.ai/llm-gateway
- **Agent CLI** — `@opperai/cli` launches Claude Code, Codex, OpenCode, OpenClaw, Hermes, Pi on Opper's gateway. https://opper.ai/agent-cli.md
- **Opper Account** — pay-as-you-go account for AI apps; users sign in with Opper, top up one balance, and pay for inference in connected apps. https://opper.ai/ai-wallet.md
- **AI control plane** — rules for every model call: spend limits, data retention, model access, checks and routing, set per organization or project. https://opper.ai/ai-control-plane
- **LLM router** — dynamic routes with fallback chains, called like any model. https://opper.ai/llm-router
- **EU-hosted models** — every model with an EU route, with the provider, country, zero data retention, training posture and price of each route. https://opper.ai/models/eu
- **Compare models side-by-side** — privacy, hosting, pricing, and benchmarks across frontier models. https://opper.ai/compare

## Why developers pick Opper

- **One key, every model.** No per-provider keys for OpenAI, Anthropic, Google, Meta, Mistral, xAI, AWS Bedrock, etc.
- **EU data residency.** Hosted in AWS Stockholm. Every route is tagged with where it runs, and the EU list is at https://opper.ai/models/eu. Calling an EU route's model id keeps that call in the EU on any plan; a Model access rule (Control Plane) allows only EU inference and storage locations, so no LLM or embedding call in the organization can reach a route outside the EU, and a call that tries fails with a 403 that names the model and the condition that failed.
- **Automatic fallbacks.** The same model wired across multiple clouds (AWS, Azure, GCP). One upstream rate-limits, the gateway routes to the next.
- **Drop-in OpenAI SDK compatibility.** Change `baseURL` and key — your code keeps working.
- **Full observability.** Every call costed and logged; full traces when tracing is on.
- **Pay-as-you-go.** No markup on tokens, 3% fee on credit purchases. Failed requests not billed. BYOK is free on the Gateway plan.

## Security and GDPR compliance

Frontier AI, EU compliance included: one DPA for every model, SSO, audit logging and model allowlists.

- **GDPR-compliant, EU-hosted.** Platform in Stockholm, EU routes end to end, and one DPA with one sub-processor for every model you add. https://opper.ai/ai-compliance
- **Zero data retention.** Opper stores no prompts or responses by default, and with ZDR on many pay-as-you-go routes by default, no content is stored end to end.
- **ISO 27001 certified.** Certified against ISO/IEC 27001:2022. AES-256 at rest, TLS in transit, controls and policies on the trust center. The certificate is available on request.
- **Plans.** Model allowlists come with the Control Plane plan; SSO/SAML and audit logging with Enterprise. https://opper.ai/pricing.md

Documents: [Data Processing Agreement](https://opper.ai/data-processing-agreement), [security overview](https://opper.ai/security-overview), [trust center](https://trust.opper.ai) (controls, policies and Opper's own sub-processor register), and the [provider directory](https://opper.ai/providers) for the model providers you route to, with where each one runs and its retention and training posture.

## Set the rules for every model call

Spend limits, data retention, model access, checks and routing, set once for the organization and tightened per project (Control Plane plan). Full reference: [ai-control-plane.md](https://opper.ai/ai-control-plane.md).

- **Spend limits.** A monthly budget for the organization and caps per project, role, user or key; the next call over the limit is refused with HTTP 402.
- **Data retention.** Opper traces kept 1 to 30 days or not at all, per project down to 0 days, plus a provider data policy of zero data retention (no training, no logging).
- **Model access.** Allow only approved providers, models, inference and storage locations (for example EU), maker or provider countries.
- **Checks.** Flag, redact or block content on every call, or score outputs after it.
- **Routing.** The provider order for pooled models: round robin by default, or cheapest, fastest, highest throughput or weighted.

## Get started

Up and running in minutes, with one key for every model and nothing to set up.

1. **Create an account.** For yourself or a company. Invite team members anytime. https://opper.ai/sign-up/free
2. **Add credits.** Free models work before you add a card. Add a card and top up credits to use premium models, or bring your own provider keys. https://opper.ai/pricing
3. **Make your first call.** OpenAI SDK compatible. Or paste the setup prompt above into your coding agent.

Sign up needs no credit card: you get an API key straight away and the free models work in the playground and the API. Add a card to use premium models, pay-as-you-go with no minimum.

## Pricing

- **No markup on tokens.** A 3% fee applies when you buy credits. No seat fees, no minimums.
- **BYOK**: bring your own provider keys, free on the Gateway plan.
- **Credits** — buy in any currency, no minimum top-up. Auto top-up via Stripe.
- Full pricing: [pricing.md](https://opper.ai/pricing.md)

## Migration

Opper is OpenAI SDK compatible. Change your base URL and API key — existing code works unchanged.

## Comparisons

- [vs OpenRouter](https://opper.ai/openrouter-alternative) — same unified API, plus self-serve EU residency on every plan, ISO/IEC 27001:2022 certification and built-in tracing.
- [vs LiteLLM](https://opper.ai/litellm-alternative) — managed, no proxy to self-host, with native tracing and EU regions.
- [vs Portkey](https://opper.ai/portkey-alternative) — the models come with the account: one key, no provider keys, no prompts stored by default, EU data residency.
- [vs Vercel AI Gateway](https://opper.ai/vercel-ai-gateway-alternative) — hosted in Europe, no markup on tokens, tracing and guardrails built in.

## Customer stories

- [Instabridge](https://opper.ai/stories/instabridge) — reliable AI agent at global scale, 200M+ downloads across 190+ countries.
- [Svenska Bostäder](https://opper.ai/stories/svenska-bostader) — AI-powered tenant services for 28k+ apartments.
- [LexBox](https://opper.ai/stories/lexbox) — litigation AI scaled to production on Opper.
- [Njord](https://opper.ai/stories/njord) — Fortune 500 deal-making for fast-growing teams across 4 continents.
- [AI-BOB](https://opper.ai/stories/aibob) — compliance reviews from days to minutes.
- [Beatly](https://opper.ai/stories/beatly) — Nordic influencer-marketing platform, agency to AI-first SaaS.

## Links

- Sign up: https://opper.ai/sign-up/free
- Docs: https://docs.opper.ai
- API reference: https://docs.opper.ai/v3-api-reference
- OpenAPI spec (JSON): https://docs.opper.ai/openapi-v3.json
- Models catalog: https://docs.opper.ai/capabilities/models
- Models hosted in the EU: https://opper.ai/models/eu
- Trust center (ISO/IEC 27001:2022): https://trust.opper.ai
- Status: https://status.opper.ai
- GitHub: https://github.com/opper-ai
- Index of agent-readable pages: https://opper.ai/llms.txt
- Full content for one-shot fetch: https://opper.ai/llms-full.txt
