The privacy-first AI coding plan.

Chat (including multimodal chat) and tools at regular API token prices. Each plan includes twice its monthly price in that usage, with a month, week and 5-hour session window. The same models and prices as the prepaid wallet.

OCR, speech, image generation, embeddings and rerank never draw from a plan. They run on the prepaid wallet: with API keys that are not on a plan, or with a plan's keys when that plan allows wallet use. The same switch lets chat and tools continue from the wallet when a plan window is full.

Usage counts against the current session (5 hours), the current week and the current month. When a window is full, requests stop until it resets, unless the plan allows prepaid wallet use in the dashboard.

No. The wallet stays for everything the plan does not cover, and as optional overage. The plan is a monthly package for the chat and tool calls coding agents make.

Yes. Plans live in a project. The project owner buys one plan per developer and assigns each to a project member. Every developer gets their own session, week and month windows. Members see only their own plan and keys and create keys for it; the owner and admins manage all plans and keys. Each plan is billed as its own subscription.

The sticker is the same number in EUR and USD (20, 100 or 200). Invoices follow your billing currency and VAT profile, like a wallet top-up. This page can show indicative VAT for your country.

An upgrade starts a new month now. You pay the new price minus unused time on the current plan. A downgrade or cancel takes effect at the next renewal.

Create an API key on your plan and set the OpenAI-compatible base URL to https://api.nodion.ai/v1. That works in OpenCode, Aider, Cline, Continue and any OpenAI SDK. Claude Code, Codex and Cursor cannot call this API. They keep their own models. Use MCP, the CLI or the platform API if those agents should deploy the app.

Chat models on this page are served in the EU. Zero data retention for all models served through the Nodion AI Gateway. The AI Inference DPA covers Nodion-hosted models.