← Back

Fair Use Policy

How Ozore keeps unlimited AI compute reliable, fast, and fair for every customer on the platform.

What “Unlimited” Means

Every Ozore plan includes unlimited AI requests. There are no per-token fees, no usage meters, and no surprise invoices. You pay a flat monthly price and use the API as much as you need.

To keep this sustainable, Ozore applies a lightweight fair use system. It exists to prevent the kind of abuse that would degrade performance for everyone (bot networks, mass scraping, denial-of-service activity), not to throttle legitimate heavy use. Sustained agentic workloads, long-running development sessions, and demanding multi-agent setups are all fully supported.

Per-Install Licensing

Each Ozore subscription covers one active AI tool environment. If you use multiple AI-powered tools (for example, Claude Code on your laptop, OpenClaw on your server, and Hermes on your phone), each environment needs its own plan and API key.

This keeps pricing simple and infrastructure allocation predictable. Every environment gets its own dedicated routing, session state, and priority tier, with no shared queues and no interference between tools. The result is faster, more reliable performance for each one.

Who These Plans Are For

Ozore consumer plans are designed for individual developers building, experimenting, and shipping with AI. Typical use cases include personal coding assistants, prototype development, side projects, learning, and daily development workflows.

You can use AI-generated output for any commercial purpose: your code, your content, your business. However, you cannot use Ozore as backend infrastructure for external users. These plans are not intended to power commercial SaaS products, customer-facing applications, or large-scale automated pipelines serving third parties. If your workload supports a commercial service or external customers, please reach out about an enterprise agreement.

Plan Tiers and Priority

Every plan gets unlimited requests. The difference between tiers is capacity: higher plans get more concurrent agent slots, larger context windows, and faster scheduling during peak demand.

  • Starter: 1 agent slot, 50K token context. Great for solo devs getting started.
  • Basic: 3 agent slots, 80K token context. Built for daily AI-powered development.
  • Pro: 5 agent slots, 100K token context. For power users running heavy agentic workflows.

Fair Resource Sharing

Ozore runs on shared AI infrastructure. During periods of high demand, higher-tier plans receive scheduling priority. That's the main reason to upgrade. Heavy legitimate use is fully supported on every tier; you won't get throttled for running long agentic sessions or working hard.

Fair use enforcement only activates against actual abuse, the kind of activity that would noticeably impact other users.

What's Not Allowed

The following activities are prohibited on all plans:

  • Sharing a single API key across multiple tool installations or users
  • Reselling Ozore access or proxying our API as your own service
  • Using the API as general-purpose compute unrelated to AI inference
  • Large-scale scraping or data harvesting where the LLM is used only to extract third-party content
  • Launching or coordinating denial-of-service attacks
  • Attempting to reverse-engineer or benchmark upstream model providers in violation of their terms
  • Generating illegal content including fraud, harassment, or CSAM

Violations may result in account suspension or termination.

Data Usage

By using Ozore, you agree that request and response data processed through the platform may be used to improve and operate the service, including but not limited to model quality evaluation, routing optimization, and training. If you prefer that your data is not used in this way, the Zero Data Retention add-on guarantees we won't train on your data and won't route it through any model that trains on inputs.

Policy Updates

This policy may evolve as the platform grows. Changes will be communicated to customers at least 1 day before taking effect.

For questions about this policy or specific use cases: [email protected]

Last updated: May 2026