Fair Use Policy
Last updated:
This Fair Use Policy (hereinafter "the Policy") describes the technical rules governing the allocation and consumption of credits on the Haloon platform. It supplements the Terms of Use, to which it is subordinate.
1. Purpose
This document describes the types of quotas that apply on the Haloon platform, depending on the chosen plan and type of usage.
2. Units of Measurement
2.1 The Credit Pool
Each plan grants a single shared pool of credits for the billing period. Every request — text, image, video or audio generation — draws from this same pool.
The number of credits a request consumes is variable and depends on the AI model used, the volume of data processed (input and output tokens), and the features activated (vision, web search, etc.). For image, video and audio generation, the credit cost is derived from the provider-reported cost of the underlying model.
The credit cost of each model is indicatively specified in the technical documentation. These values may change based on third-party provider pricing (see section 3.3 of the Terms of Use).
2.2 File Size
The maximum size of files that can be submitted to the platform (documents, images, etc.) is limited depending on your plan, expressed in megabytes (MB). Files exceeding this limit are rejected at the time of submission.
3. Limits per Plan
| Plan | Credit pool | Max file size |
|---|---|---|
| Free | 20 000 000 credits | 0 MB |
| Discovery | 300 000 000 credits | 5 MB |
| Expert | 900 000 000 credits | 20 MB |
| Max | 3 000 000 000 credits | 20 MB |
Standard plans cover the needs of 90% of users. Plans with custom quotas (volume, enterprise) are subject to a specific commercial agreement.
4. Behavior upon Exhausting the Credit Pool
When your credit pool is exhausted:
- Premium models and image, video and audio generation become unavailable until the pool renews at the start of the next billing period.
- Generic (unlimited) text models remain available and keep working normally.
4.1 Model Types
Unlimited models are:
- ChatGPT Instant
- DeepSeek
- Gemini Flash
- Llama OpenSource
- Llama - Fast
- Llama - Thinking
- Magistral
- Mistral Instant
All other models are considered premium models and are subject to quotas.
5. Special Cases
5.1 Large Requests
Certain operations (long document analysis, high-resolution image generation, conversations with extended context) may consume a significant number of credits in a single request. It is your responsibility to anticipate this consumption before submitting such requests.
5.2 API Access
API access is subject to a specific commercial agreement that defines its own quotas and credits.
6. Transparency and Monitoring
You can view your current consumption and the time remaining before renewal at any time from your account space, under the Subscription section.
7. Changes to this Policy
Haloon reserves the right to modify the technical parameters of this Policy (credit costs, per-plan limits, etc.) in order to ensure the quality of the Service.
Modifications resulting in a significant reduction of the available quota will be communicated in advance by email or in-app notification. Minor adjustments or improvements may be applied without prior notice.
8. Contact
For any questions regarding your quota or this Policy:
- Email: contact@haloon.ai
- Account space: Help & Support section