Terms of Service

By creating an account, generating an API key, sending HTTP requests to the Model Mesh gateway (https://www.modelmesh.in/api/v1), or accessing the Model Mesh dashboard, you (“Customer”, “Developer”, or “User”) agree to be legally bound by these Terms of Service (“Terms”).

If you are entering into these Terms on behalf of an entity, corporation, or organization, you represent and warrant that you possess the requisite authority to bind that entity to these Terms. If you do not have such authority or do not agree with any provision, you must refrain from accessing or using the Model Mesh platform.

Model Mesh operates a low-latency, unified AI routing infrastructure that standardizes access to leading foundation models (including OpenAI, Anthropic, Google, Groq, Together, DeepSeek, and open-source clusters) through a single OpenAI-compatible API interface.

Model Mesh reserves the right to modify, optimize, expand, or deprecate specific model checkpoints, upstream clusters, or fallback paths to maintain overall network stability and security.

To access the platform, you must register using a verified email address or approved OAuth provider. You are solely responsible for maintaining the confidentiality of your credentials and any Model Mesh API keys (prefixed with mm_live_ or cc_) generated under your account.

• Key Confidentiality: You must never expose your API keys in public repositories, client-side bundles, browser code, or shared environments.
• Compromise Notification: If you suspect that an API key has been exposed or compromised, you must immediately revoke it via the Model Mesh dashboard and issue a replacement.
• Account Responsibility: All API requests, token charges, and actions initiated using your credentials or API keys are deemed to be authorized by you.

Model Mesh utilizes an exact micro-cent metering engine (where 1 USD is divided into 100,000,000 micro-units). Usage is calculated per token (prompt and completion tokens) based on published rates in our Model Catalog.

Users who voluntarily contribute third-party provider keys (e.g. OpenAI, Anthropic, Groq, Together) to the Model Mesh Community Pool earn pool credits and fair-share allowances according to our current reward formulas.

• Encryption at Rest: All contributed upstream keys are encrypted immediately using AES-256-GCM master keys and are never displayed in plaintext or returned through the API.
• Revocation: You retain full ownership of your contributed keys and may revoke or delete them from the Model Mesh dashboard at any time. Revocation takes effect across router memory in under 60 seconds.
• Legitimate Keys Only: You warrant that you have full legal rights to use and contribute any key submitted to the pool and that contributing the key does not violate your agreement with the respective upstream provider.

Your Content: As between you and Model Mesh, you retain all rights, title, and interest in and to the data, prompts, images, video references, code snippets, and instructions you submit to the Model Mesh API (“Input”), as well as the resulting completions and responses generated by the upstream models (“Output”).

No Training on Prompts: Model Mesh does not use your Inputs or Outputs to train, fine-tune, or improve any artificial intelligence or machine learning models. Your data is passed transiently to the upstream provider strictly for inference execution.

Model Mesh targets an operational availability of 99.9% for our Direct Upstream API endpoints. Our multi-cloud edge infrastructure continuously monitors upstream provider health, automatically re-routing requests if a provider node reports degraded response times or HTTP 5xx errors.

Planned maintenance windows will be announced via our Status Page and Changelog with reasonable advance notice. Unplanned outages or degradation caused by upstream foundational providers (e.g., global OpenAI or AWS cloud outages) are mitigated by our fallback routing heuristics where technically feasible.

To safeguard network throughput and fair access, Model Mesh applies request-per-minute (RPM), token-per-minute (TPM), and concurrent connection limits based on your account plan.

When limits are reached, the API returns an HTTP 429 (“Too Many Requests”) response accompanied by standardRetry-After headers. Any attempt to circumvent rate limits through distributed botnets, artificial key rotation, or denial-of-service tactics constitutes a material breach of these Terms.

THE MODEL MESH API AND ALL ASSOCIATED SERVICES ARE PROVIDED ON AN “AS IS” AND “AS AVAILABLE” BASIS WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED. MODEL MESH DISCLAIMS ALL WARRANTIES, INCLUDING MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE, AND NON-INFRINGEMENT.

IN NO EVENT SHALL MODEL MESH, ITS DIRECTORS, EMPLOYEES, OR AGENTS BE LIABLE FOR ANY INDIRECT, CONSEQUENTIAL, INCIDENTAL, PUNITIVE, OR SPECIAL DAMAGES, OR FOR LOSS OF PROFITS, DATA, OR BUSINESS INTERRUPTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR INABILITY TO USE THE SERVICE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

These Terms and any dispute arising out of or related to them shall be governed by and construed in accordance with applicable laws, without giving effect to any conflict of law principles.

The parties agree to first attempt informal resolution for any dispute by notifying the other party in writing. If the dispute is not resolved within 30 days of notice, it shall be submitted to confidential, binding arbitration conducted in the English language.

If you have questions, notices, or require formal correspondence regarding these Terms of Service, please reach out to: