Skip to content

One API for your agents.

Open models, web and social search, quick decisions and company lookups for your agents. One key, billed by use, fully managed in the EU.

Create a keyBook a call

Included in Nurion Work, Meet and Desk. After that, you pay per call from prepaid credit.

Every endpoint, one key.

Point your agent at api.nurion.com. It speaks the OpenAI and the Anthropic format, so the tools you use today connect as they are.

ModelsChat, vision and tool calls on open models/v1/chat/completions
Price: Per token, by model
Included per person and month: Work€5Meet€10Desk€50
DecisionsA choice, a score or a yes or no, with probabilities/v1/systemone
Price: About €0.0002 a question
Included per person and month: Uses the Models allowance
Web searchCurrent results from the web, with sources/v1/search
Price: €0.02 a search
Included per person and month: Work100Meet200Desk500
Social searchPosts from X, TikTok, Instagram and Bluesky/v1/social-search
Price: €0.02 a search
Included per person and month: Work50Meet100Desk300
Company lookupsCommercial register facts about any company/v1/company-lookup
Price: €0.20 a lookup
Included per person and month: Work3Meet10Desk50

Ready for your agent.

Works with OpenCode, Pi, Claude Code and anything that speaks the OpenAI or Anthropic format. Paste the config, and we run everything else.

opencode.json

{
  "provider": {
    "nurion": {
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "https://api.nurion.com/v1",
        "apiKey": "{env:NURION_API_KEY}"
      },
      "models": { "nurion-qwen38": {} }
    }
  }
}

Your prompts stay yours.

Models run on our own GPUs in EU data centers. Prompts and answers are not logged.

Each model has a per-minute limit, so no one fills a GPU alone. Need more? Run a model of your own. Private Inference

Give your agents one key.

Included in Nurion Work, Meet and Desk. After that, you pay per call from prepaid credit.

Create a keyBook a call