Skip to content
Navorynth
  • How it works
  • Build
  • Founding Beta
  • Security
Join the Founding Beta
  • How it works
  • Build
  • Founding Beta
  • Security
Join the Founding Beta

DRAFT — FINAL TERMS WILL BE AVAILABLE BEFORE PURCHASE

This document is a working draft. It has not been reviewed or approved by a lawyer, is not legal advice and is not yet in force. Bracketed items are still to be completed.

Document status: DRAFT — REQUIRES LEGAL REVIEW

Not reviewed by a lawyer. Not legal advice. The technical statements are taken from NETWORK.md and DATA_AND_PRIVACY.md; provider terms are summarized, not restated: read the providers' own documents.

Source: Navorynth AI Provider Disclosure (draft), customer/AI_PROVIDER_DISCLOSURE.md

Your AI providers: bring your own key (DRAFT)

DRAFT — REQUIRES LEGAL REVIEW.

Navorynth does not include any AI model. It uses your own accounts with the AI providers you connect (OpenAI, Anthropic, Google Gemini, xAI), and they bill you directly. Building software needs OpenAI and Anthropic.

Which providers, and what they do

Provider Role in Navorynth How it is reached
OpenAI Product Manager (plans the product) and Reviewer (checks every step, including screenshots) Navorynth calls the OpenAI API directly from your computer with your OpenAI API key
Anthropic Builder (writes the code and runs the tests) Navorynth runs Claude Code, Anthropic’s program that you install yourself, with your Anthropic API key

You need an API account with each provider, with credit or a payment method: platform.openai.com and console.anthropic.com. A ChatGPT or Claude subscription is not an API key and cannot be used.

What is sent, and when

Only when a model is called: while planning, building and reviewing. Opening Navorynth, browsing projects, the license check and the demo send nothing.

  • To OpenAI: your idea and answers, the plan, excerpts of project files you sent to the team, the Builder’s proposed actions, filtered code changes, test output and screenshots of your product.
  • To Anthropic (through Claude Code): instructions, project rules, file excerpts, the code the Builder reads and writes, and the results of its commands.
  • Key checks: at setup and before work starts, each key is checked with its provider (no AI tokens used).

Everything goes directly from your computer to the provider. There is no Navorynth server in between, and the seller never sees your keys, prompts or code. Known secret files are excluded and credential-like text is redacted before sending, but you decide what goes into a project. See DATA_FLOW.md.

Who bills you

  • OpenAI and Anthropic charge your accounts for usage, at their prices. Navorynth does not charge for, resell or intermediate AI usage, and the seller cannot refund it.
  • Navorynth shows estimated costs and enforces the maximum budget you set for each project, between model calls. A request already in progress can overshoot it slightly, prices can change and provider billing can lag: the provider’s bill is what you owe.
  • Set spending limits in both provider accounts as a second safety net.

Provider terms apply

Your use of each provider is governed by your agreement with that provider, including its usage policy, privacy terms and data-retention settings, for example:

  • OpenAI: Services Agreement or the terms that apply to your account, and Usage policies.
  • Anthropic: Commercial Terms or the terms that apply to your account, and the Usage Policy; Claude Code is Anthropic software under Anthropic’s terms.

How long the providers keep your data, and whether they use it, is decided by those terms and your account settings, not by Navorynth. Navorynth sends OpenAI requests with store=false. [REVIEW: which provider terms bind individual vs. business customers.]

Your responsibilities

  • Keep your accounts in good standing and your keys secret. Never share your keys, including with us: we will never ask for them.
  • If you think a key was exposed, revoke it in the provider’s console and enter a new one in Navorynth.
  • Follow the providers’ usage policies (ACCEPTABLE_USE.md).
  • Models can be changed, priced differently or retired by the providers. Navorynth cannot control this.

How Navorynth treats your keys and Claude Code

  • Keys are stored in Windows Credential Manager on your computer and sent only to the provider they belong to. They are never written to files or logs and never enter the build sandbox.
  • Navorynth runs Claude Code unmodified. In its own sessions it uses your API key only and turns off Claude Code’s non-essential traffic (telemetry, error reporting, auto-update). Your own use of Claude Code outside Navorynth is between you and Anthropic.

Names

OpenAI and ChatGPT are trademarks of OpenAI. Anthropic, Claude and Claude Code are trademarks of Anthropic. Navorynth uses their models and software through your accounts; it is not affiliated with, sponsored or endorsed by either company. [REVIEW]

Navorynth

Your AI team for work that takes more than one prompt.

A desktop app for Windows 10 and 11. Founding Beta, by request.

  • How it works
  • Founding Beta
  • Security and privacy
  • Get Navorynth
  • Questions
  • Privacy (draft)
  • Terms (draft)
  • AI providers (draft)

© 2026 Navorynth. All rights reserved.

OpenAI, Anthropic, Google Gemini and xAI are AI providers you can connect with your own API keys. Navorynth is not affiliated with or endorsed by any of them.