TERMS & CONDITIONS

01 About These Terms & About Mesh

These Terms & Conditions (the "Terms") form a binding agreement between you and Mesh Labs Cloud, Inc., a Delaware corporation having its mailing address at 2810 North Church Street, Wilmington, Delaware 19802, United States ("Mesh", "we", "us" or "our"). The Terms govern your access to and use of MESH, including the meshlabs.cloud website, the MESH control plane and APIs, the MESH browser-based grid node, the MESH desktop application, the MESH mobile application, the MESH IDE extensions (for Visual Studio Code, Cursor, Antigravity and compatible editors), the MESH command-line and SDK clients, and any related software, services, documentation, samples and content (collectively, the "Service").

By creating an account, connecting a node, submitting a job, or otherwise accessing the Service, you confirm that you have read, understood and agree to be bound by these Terms and by our Privacy Policy. If you do not agree, you must not use the Service.

If you are accepting these Terms on behalf of an organization, you represent that you have the authority to bind that organization, and "you" refers to both you individually and that organization.

02 The MESH Service

In plain terms: MESH lets people and organizations run artificial-intelligence tasks using computing capacity supplied by participating devices. Instead of relying only on a single data center, MESH can divide work among laptops, desktops, servers, mobile devices and other compatible hardware. If you contribute a device to the open Global network and its work is verified, you may earn Compute Units ("CU"). MESH includes systems that sign users in, schedule and validate work, connect devices directly, cache large AI model files on devices, and let developers use MESH through apps, extensions, SDKs and APIs.

MESH is a decentralized AI compute network. The Service orchestrates artificial-intelligence inference (and certain training-class) workloads across a dynamic mesh of heterogeneous client devices acting as compute nodes. Workloads may be partitioned into shards, distributed peer-to-peer (including via WebRTC and torrent-style swarming), executed on browser or native backends such as WebGPU, WebNN, WebAssembly and CPU/GPU runtimes, validated through consensus and integrity mechanisms, and rewarded through a contribution-tracking system referred to as "Compute Units" or "CU" (see Section 9).

The Service is composed of, among other things:

• a hosted control plane that handles authentication, job intake, scheduling and accounting;
• a browser-based grid node application, an installable desktop node application, and IDE extensions that execute shards using your device's CPU/GPU/NPU;
• peer-to-peer transport mechanisms that allow nodes to exchange shards, tensors and results;
• persistent local storage on your device used to cache model assets (in the browser, the Origin Private File System / "OPFS"; in the desktop application, a cache directory within the application's data folder on your file system; in the IDE extensions, the editor's extension storage directory);
• SDKs, APIs and reference clients for embedding the Service in third-party applications.

03 Eligibility & Account Registration

To use the Service, you must be at least 18 years old (or the age of legal majority in your jurisdiction, whichever is greater), have the legal capacity to enter into a binding contract, and not be located in, organized under the laws of, or ordinarily resident in a jurisdiction subject to comprehensive sanctions by the United States, the European Union, the United Kingdom, Israel or the United Nations Security Council. You may not use the Service if you are listed on any sanctions list or denied-parties list.

Some features require an account. You agree to provide accurate, complete and current information at registration and to keep that information up to date. You are solely responsible for keeping your credentials, API keys and node identifiers confidential and for all activity occurring under your account. You must notify us promptly at security@meshlabs.cloud if you suspect unauthorized use.

04 Roles: Requesters, Providers & Hybrid Nodes

The Service supports two primary roles, plus a hybrid mode:

• Requester (Demand side). Submits AI inference or training jobs to the network and consumes results.
• Provider (Supply side). Contributes idle compute capacity by allowing a browser-based, desktop, mobile or IDE node to load model shards and execute inference tasks on the Provider's device. Verified contributions to the open Global network may earn Compute Units (CU); contributions made while a private cluster is the active compute target are donated to that cluster and do not earn personal CU — see Section 9.3.
• Hybrid Node. Acts as both Requester and Provider, with role assignment chosen by the user or selected automatically based on device and network state.

You choose your role and may switch between roles at any time, subject to the operational policies of the Service. Different sections of these Terms apply depending on the role you take on at any given moment.

05 Acceptable Use

You agree not to, and not to permit any third party to:

• use the Service to generate or distribute content that is unlawful, infringing, defamatory, obscene, sexually explicit involving minors (CSAM), incitement to violence, harassment or hate speech, or that violates the privacy or publicity rights of others;
• generate non-consensual intimate imagery, deceptive deepfakes intended to defraud, or content reasonably likely to cause physical harm;
• use the Service to develop, train or deploy systems for biometric identification, mass surveillance, social scoring, or other uses prohibited under applicable law (including the EU AI Act where applicable);
• use the Service to design, manufacture or operate weapons (including chemical, biological, radiological or nuclear weapons), critical-infrastructure attacks, malicious code or cyber-weapons;
• circumvent, probe or interfere with the security, integrity, validation, reputation or consensus mechanisms of the Service, including submitting falsified compute results, spoofing node identity, conducting Sybil attacks, tampering with hardware attestations, or poisoning tensor outputs;
• introduce malware, exfiltrate data through the peer network, or use peer connections for any purpose unrelated to executing the assigned MESH workload;
• reverse-engineer, decompile or disassemble the Service except to the extent expressly permitted by mandatory applicable law;
• scrape, crawl, mirror or systematically copy the Service, the model registry or any outputs except as expressly permitted by us in writing;
• resell, sublicense or commercially exploit the Service or Compute Units (CU) other than as expressly permitted by these Terms;
• submit, process or store through the Service any data subject to heightened regulatory requirements, including protected health information under HIPAA, nonpublic personal information under GLBA, sensitive personal data (including biometric, genetic, health, sexual-life, racial or ethnic-origin, political, religious, philosophical or trade-union data), children's personal information subject to COPPA, payment-card data subject to PCI DSS, government identification numbers, or other data subject to special security, privacy or handling requirements, unless we have expressly agreed in writing that the applicable Service configuration supports that data;
• where you use a network execution mode, submit trade secrets, credentials, confidential source code or other confidential information unless you are authorized to disclose it to the Provider nodes that may process it; or
• violate any applicable law, third-party right or these Terms.

Without limiting the foregoing, Mesh may, in its discretion and without prior notice, investigate suspected violations or security threats; preserve and disclose logs, telemetry, metadata and other records as reasonably necessary for the investigation or as required by law; suspend or terminate any job, shard, node session or peer connection while the investigation is pending; and take reasonable steps to protect the integrity, security and availability of the Service and its users.

06 Provider-Specific Terms

When you operate a node as a Provider, you grant Mesh and Requesters whose jobs are routed to your node a limited, non-exclusive, royalty-free, worldwide right to:

• load, execute and unload model shards on your device using your device's available compute backends (e.g., WebGPU/WebNN/WASM/CPU in the browser, and equivalent native acceleration in the desktop application);
• persistently cache shards, manifests and auxiliary assets in local storage on your device (including browser OPFS and, in the desktop application, a cache directory on your file system, which you can remove by clearing the cache in the application or uninstalling it);
• establish peer connections (including WebRTC, with or without TURN relays) and exchange shards, tensors, intermediate states, signaling and telemetry with other peers and with the control plane;
• profile your device's compute capabilities, monitor heartbeats and lifecycle events, and record contribution metrics for accounting and reputation purposes;
• where you run the desktop application or an IDE extension, read your device's name (hostname), system resource state (memory, GPU/VRAM, CPU load, disk capacity), power and thermal state, and screen-lock / user-activity state, solely to identify your devices in your dashboard, schedule work when the device is idle, and yield resources to you as described in Section 6.1; and
• cryptographically attest your device identity (where supported) to mitigate Sybil attacks, as further described in the Privacy Policy.

6.1 Device Capacity, Power, Thermal & Data Usage

Operating a Provider node consumes electricity, produces heat, and may transfer significant volumes of network data — including downloading model files that can reach multiple gigabytes, and uploading model shards, intermediate tensors and results to other peers as part of the Service's peer-to-peer distribution. You are solely responsible for any data-cap, overage or bandwidth charges imposed by your internet service provider or mobile carrier. The Service includes a deterministic Resource Arbitration Engine reasonably designed to yield to the host operating system, scale down parallelism on thermal events, and respect device-sleep and battery-saver settings, but you remain solely responsible for monitoring your device, its operating environment, and the cost and consequences of the electricity it consumes. Do not operate a Provider node on hardware you do not own or are not authorized to use, on metered or shared-quota connections without consent, or in environments where elevated GPU utilization could cause damage.

6.2 No Obligation; Right to Disconnect

Providers are not employees, contractors or agents of Mesh. You may disconnect your node at any time, for any reason. We may, however, expire any leases on uncompleted work, withhold CU for the affected work units, and adjust your reputation accordingly. All job routing and shard assignment is performed automatically by the Service's scheduling systems in our discretion. You cannot preview, select, filter, approve or reject particular jobs or shards before assignment. You may not see the Requester's identity, the purpose of the job, or input content beyond what your device must receive to perform the assigned computation. You acknowledge that you cannot control which jobs are routed to your node and that we may change scheduling criteria and routing logic at any time.

6.3 Provider Outputs & Confidentiality of Inputs

Inference inputs (including prompts, source images, source video, and intermediate tensors), and any associated outputs, assigned to your node are the confidential property of the requesting party. You agree not to retain, copy, log, redistribute or use such inputs or outputs other than as required to execute the assigned shard, and not to attempt to deanonymize or correlate inputs or outputs across jobs. You agree to participate in the Service's trustless validation protocol (including canonicalization, probabilistic hashing, redundant execution and quorum acceptance), and you understand that divergent or fraudulent results may result in reputation penalties, withholding of CU, or account suspension.

6.4 Reputation System

Mesh maintains an internal reliability index for each Provider node ("Reputation"). Reputation is an operational measure of node reliability, performance and trustworthiness. It is based on technical and operational factors relating to the node, not personal characteristics about its operator. Relevant factors may include successful completion and validation of assigned work, result accuracy, uptime, availability, response latency, disconnections, incomplete work, participation in validation, and anomalous, divergent or potentially fraudulent activity.

Reputation may affect the volume, priority and type of work routed to a node. Higher-Reputation nodes may receive more frequent or higher-value assignments; lower-Reputation nodes may receive fewer assignments, be deprioritized, or be temporarily or permanently excluded from assignments.

We determine the methodology, weighting, thresholds and operational consequences of Reputation and may modify them as the Service evolves. Scores may not always be visible. Where a decision based primarily on automated processing produces legal or similarly significant effects, any rights to an explanation, human review or challenge provided by applicable law remain available as described in the Privacy Policy.

07 Requester-Specific Terms

When you submit jobs as a Requester, you understand and agree that:

• your inputs may be processed on Provider nodes that are not under Mesh's control and that are by design treated as untrusted (this is the basis of the trustless validation protocol). You should not transmit personal data, regulated data (e.g., PHI, financial data subject to PCI DSS) or trade secrets into the Service unless and until you have determined that the Service's privacy and validation guarantees are adequate for your use case. Unless we expressly agree otherwise in writing, we make no representation that the Service is configured to comply with HIPAA, GLBA, PCI DSS or another sector-specific regulatory regime;
• the Service's validation, redundancy, hybrid-escalation and checkpoint/resume mechanisms reduce, but do not eliminate, the risk that a result is incorrect, incomplete or non-deterministic;
• you are solely responsible for the inputs you submit and for any outputs you generate, store, distribute or commercialize, including for compliance with applicable law and non-infringement of third-party rights (including copyright, trademark, publicity, privacy and AI-specific regulations);
• generated outputs may be substantially similar to outputs generated by other Requesters using the same or similar inputs and models. We do not guarantee uniqueness of outputs.

08 Organization Clusters

In addition to the Global network and personal or invitation-based private clusters, the Service may offer organization-owned collaborative clusters ("Organization Clusters"). An Organization Cluster allows an organization to submit workloads for execution by eligible devices associated with that cluster. Those devices may be owned or operated by the organization, its personnel, its members, or third-party suppliers that an authorized cluster administrator has invited or approved.

8.1 How Organization Clusters Work

Workloads submitted to an Organization Cluster are scheduled to eligible nodes participating in that cluster and are not submitted to the open Global network as Global-network workloads. However, an Organization Cluster is not an entirely offline or device-only service. Mesh-controlled infrastructure may process authentication, job intake, scheduling, signaling, validation, billing and technical metadata, and job inputs, intermediate artifacts and outputs may pass through or be temporarily stored in cloud infrastructure and S3-compatible object storage. Download links may be time-limited bearer links that can be used by anyone who obtains the complete link until it expires.

A supplier whose active compute target is an Organization Cluster may separately enable fallback of its idle device to Global-network work when the cluster has no eligible work. That setting does not convert an Organization Cluster workload into a Global-network workload, but it means the supplier's device may also process unrelated Global network jobs at other times.

8.2 Administrator Responsibilities

An organization must designate one or more authorized administrators to manage cluster membership, device participation, personnel access, workload submission and available routing settings. The organization is responsible for ensuring that its administrators, members and suppliers are authorized to participate; for the security, configuration, retention and disposal practices of devices and systems under its control; and for establishing a lawful basis and providing required notices for personal data it submits.

Mesh remains responsible for the Mesh-controlled portions of the Service as described in these Terms, the Privacy Policy and any separately applicable data processing agreement. Organization Cluster contributions do not accrue personal CU, as described in Section 9.3. Fees, service levels, eligibility requirements and other commercial terms for an Organization Cluster may be set out in an order form, subscription plan or other written agreement.

09 Compute Units, Credits & Payments

9.1 Compute Units; Provider Payment Claims

"Compute Units" ("CU") are an internal, non-transferable accounting unit used to display verified compute contributions and the cost of submitted jobs. CU are displayed with the "⚡ CU" designation; references in older Service materials to "Karma" mean CU.

Verified Provider earnings accrue as a US-dollar-denominated contractual payment obligation of Mesh to the eligible Provider and may be displayed in CU at the conversion rate published in the Service (currently 100 CU = US$1.00). A future rate change will apply prospectively and will not reduce a US-dollar Provider earning that has already accrued. Subject to applicable law, identity and tax verification, sanctions screening, published minimum thresholds, fraud review, fees and processing times, an eligible Provider may request payment of its available accrued Provider earnings using a supported payout method. Payments are made through third-party payment providers, such as Stripe or another appropriately regulated provider, and may be delivered by bank transfer or, where offered and lawful, a supported dollar-denominated stablecoin.

CU cannot be sold, assigned, transferred, traded, exchanged, gifted, pledged, encumbered or moved between accounts, and no marketplace for CU is provided or permitted. CU do not represent customer funds held on deposit or in custody. Except for Mesh's contractual obligation to pay eligible accrued Provider earnings as described above, CU are not money, electronic money, stored value, a security, an investment contract, a payment instrument or a digital asset, and have no value outside the Service. Availability and treatment of CU and Provider payouts may differ or be restricted by jurisdiction.

9.2 Reward Calibration

The amount of CU earned for a completed task is determined automatically. We estimate the computing effort required by running or profiling the workload on a standard reference device (the "Oracle Node") and consider factors such as accelerator and CPU work, data transferred, storage used, successful validation and other infrastructure costs. This is called reward calibration. We may update the formula and weights prospectively, and users cannot customize them.

9.3 Private Clusters vs Global Network Earnings

The Service distinguishes between the open Global network (the spot market for compute) and private clusters (including a personal default cluster created for your account and clusters you create or join by invitation). Your active compute target determines where your contributed hardware is offered.

While a private cluster is your active compute target, compute you contribute is donated to that cluster's workloads and does not accrue personal CU to your Provider balance for earnings or withdrawal purposes. Personal CU for contributing spare compute accrues only while your active compute target is the Global network. Switching targets does not retroactively credit prior private-cluster contributions. Metrics labeled as "contributed" or "donated" CU on a cluster page measure capacity given to that cluster and are not a personal earnings balance. Jobs you request on a private cluster may be subject to separate billing rules described in the Service.

9.4 Forfeiture & Adjustments

We may adjust, withhold, recall or cancel CU where we reasonably determine that they were obtained through fraud, error, validation failure, breach of these Terms, sanctions evasion, or violation of law. CU earned by a Provider may be reduced if the Provider's results fail validation or its node is implicated in circumvention, a Sybil attack or similar misconduct. Adjustments will not affect an undisputed Provider earning that has already been paid.

9.5 Taxes

You are solely responsible for determining and paying any taxes (including value-added, income, capital-gains and withholding taxes) arising from your use of the Service, use of CU, or receipt of Provider earnings. Where required by law, we may collect and remit applicable taxes and may request tax-residency information from you.

10 Intellectual Property & Generated Outputs

The Service, including all software, code, models authored or fine-tuned by Mesh, manifests, documentation, designs, trademarks, logos, inventions and other intellectual property, are owned by Mesh, its affiliates or its licensors and are protected by intellectual-property laws. We grant you a personal, non-exclusive, non-transferable, revocable license to use the Service in accordance with these Terms and any applicable open-source licenses for components we make available under such licenses.

As between you and Mesh, you as a Requester retain ownership of inputs you submit. To the extent generated outputs are eligible for protection under applicable law, we assign to you any rights we may have in such outputs, subject to: (i) the rights of the underlying model authors and any applicable model license, and (ii) our right to use de-identified, aggregated metadata about jobs and outputs to operate, secure, evaluate and improve the Service.

For the avoidance of doubt, nothing in these Terms grants a Requester any right, title or interest in or to a Provider's device, software or other property.

You acknowledge that AI-generated outputs may not be eligible for copyright protection in some jurisdictions and that the Service relies on third-party model weights subject to their own licenses (e.g., Stable Diffusion, FLUX, LTX-Video, ESRGAN, BiRefNet, Hunyuan3D and others). You are responsible for compliance with those underlying model licenses.

11 Data & Privacy

Our processing of personal data is described in our Privacy Policy, which forms part of these Terms. By using the Service you acknowledge that personal data (including, in some configurations, IP addresses, peer identifiers, device fingerprints, telemetry and prompt content) may be processed within the Service in accordance with that policy.

If you submit personal data through an Organization Cluster, you are responsible for determining the purposes and lawful basis for that submission and for providing required notices to affected individuals. Mesh processes organization workload content to provide the requested Service, subject to any separately applicable data processing agreement. Mesh may separately act as controller or business for account administration, security, fraud prevention, billing, telemetry, reputation and compliance processing, as described in the Privacy Policy.

If you submit personal data to the Global network, you must have a lawful basis and all rights necessary to disclose it to Mesh and the Provider nodes assigned to the job. Provider nodes are subject to the confidentiality and use restrictions in Section 6.3, but they are untrusted devices outside Mesh's physical control.

12 Third-Party Models, Software & Networks

The Service uses, integrates with, or relays through third-party components, including but not limited to: ONNX Runtime Web, WebGPU/WebNN/WebAssembly browser APIs, WebRTC, public STUN/TURN servers, signaling infrastructure, third-party model weights, model registries and hosting services (e.g., Hugging Face, from which some clients download model weights directly), public package registries, payment processors, and cloud hosting providers (including AWS). Your use of those components is also subject to their respective terms, and we make no representation or warranty regarding their availability, security or compliance.

13 Agentic Features & Actions on Your Behalf

13.1 What Agentic Features May Do

If you enable agentic functionality in a MESH desktop application or another supported client ("Agentic Features"), an AI model and connected tools may perform operations on your device or through accounts and services you connect. An "Agent Action" includes any operation performed by an automated model or tool through the Service. Depending on the capabilities and permissions you enable, Agent Actions may:

• inspect directories and read, create, rename, move, edit, overwrite or delete local files;
• execute commands, scripts or other software on your device;
• operate an in-app browser, navigate websites, enter information, click controls, upload files and submit forms;
• create, publish, upload, comment on or otherwise distribute content on the internet in your name;
• draft, address and send messages, emails or other communications in your name;
• invoke third-party APIs, MCP servers, skills, applications and connected services; and
• where expressly enabled, use paid services, make purchases or otherwise cause fees or charges to be incurred.

13.2 Your Authority, Permissions & Responsibility

You authorize the Service to perform Agent Actions only within the permissions, connected accounts, approved actions and bounded unattended workflows that you or an authorized administrator configure. You represent and warrant that you have all rights, permissions and authority necessary to access and act upon each device, file, folder, account, credential, recipient, payment method and third-party service you make available to an Agentic Feature. Agentic Features are not available to anyone under eighteen (18).

You are responsible for your instructions, capability settings, approvals, connected accounts and use of completed Agent Actions. You must review proposed actions and outputs as appropriate, maintain independent backups of important data, apply the least privileges needed, and supervise consequential activity. If you authorize an ongoing or unattended workflow, the Service may take actions within that authorization without asking for separate approval each time. Revoking a permission or disconnecting an account prevents future use after the revocation takes effect but may not stop or reverse an action already submitted to a device or third-party service.

13.3 Connected Accounts, Browsers, MCP Servers & Skills

Connected websites, APIs, MCP servers, skills and other integrations may be operated by third parties and may receive your prompts, files, browser content, tool inputs or other information when you direct the Service to use them. They are governed by their own terms, privacy notices, permissions, rate limits and automation rules. Unless we expressly state otherwise, Mesh does not operate, endorse or control them and is not responsible for their acts, omissions, security, availability or use of information.

You must not use Agentic Features to harvest credentials; gain unauthorized access; bypass CAPTCHAs, access controls, rate limits or platform restrictions; impersonate another person; engage in phishing, spam, unlawful scraping, deceptive automation or fake engagement; contact people unlawfully; or make an unauthorized purchase or transaction. Banking, healthcare, government, legal, employment and other sensitive services may require human takeover or may be unavailable for autonomous operation.

13.4 Assumption of Risk

To the maximum extent permitted by applicable law, you assume the risks arising from Agent Actions that you request, approve or permit through your configured permissions. Mesh does not guarantee that an agent will correctly understand your intent, choose an appropriate action, comply with a third party's rules, preserve data, avoid disclosure, or achieve the expected result. The disclaimers and limitations in Sections 15 and 16 apply to Agentic Features, without limiting any liability that cannot lawfully be excluded.

The terms "agent" and "Agentic Features" describe software functionality only. They do not create an employment, fiduciary, partnership or legal-agency relationship between you and Mesh.

14 Beta, Proof-of-Concept & Experimental Features

Some functionality is offered as alpha, beta, preview, proof-of-concept or experimental ("Pre-Release Features"). Pre-Release Features may be modified, throttled, removed or replaced at any time without notice; may produce inconsistent results or none at all; may not have completed security review; and are excluded from any service- level commitments. Use of Pre-Release Features is at your sole risk.

15 Disclaimers & "AS-IS" Service

To the maximum extent permitted by applicable law:

• the Service is provided "AS IS" and "AS AVAILABLE", with all faults and without warranty of any kind, whether express, implied, statutory or otherwise, including any implied warranties of merchantability, fitness for a particular purpose, title, non-infringement, accuracy of outputs, or availability;
• we do not warrant that the Service will be uninterrupted, timely, secure or error-free, that defects will be corrected, that any peer or node will be available, or that outputs will be accurate, lawful, suitable or non-infringing for your intended use;
• peer-to-peer execution involves untrusted devices and networks and may produce non-deterministic, erroneous, or maliciously crafted results despite our reasonable validation measures;
• operating a Provider node may, in some configurations, cause heat, fan noise, accelerated wear of components, increased electricity consumption, significant network data usage (download and upload), and reduced device responsiveness. You assume those risks.
• Agentic Features may make incorrect or unintended changes, delete or corrupt files, expose confidential or personal information, transmit communications or publish content in your name, invoke third-party services, incur charges, or cause third parties to restrict or terminate your accounts. You are responsible for reviewing, supervising and backing up data affected by Agent Actions.

Some jurisdictions do not allow the exclusion of certain warranties; in those jurisdictions, such warranties apply only to the minimum extent required by mandatory law.

16 Limitation of Liability

To the maximum extent permitted by applicable law, in no event will Mesh, its affiliates, officers, directors, employees, agents, licensors or service providers be liable to you for any indirect, incidental, special, consequential, exemplary or punitive damages, including lost profits, lost revenue, lost data, business interruption, loss of goodwill, procurement of substitute services, hardware damage, electricity, data-transfer or connectivity costs, file deletion, modification or corruption, unauthorized or unintended disclosure, messages, emails, publications, uploads, purchases, fees, account restrictions, third-party claims, or other consequences of Agent Actions or peer behavior, however caused and under any theory of liability, even if Mesh has been advised of the possibility of such damages.

Our aggregate liability arising out of or relating to the Service or these Terms will not exceed the greater of: (a) the total amounts paid by you to Mesh for the Service in the twelve (12) months immediately preceding the event giving rise to the claim, or (b) one hundred US dollars (US$100). Multiple claims do not enlarge this cap.

Nothing in these Terms excludes or limits any liability that cannot be excluded or limited under applicable law (including liability for fraud, gross negligence, willful misconduct, or death or personal injury caused by negligence).

17 Indemnification

You will defend, indemnify and hold harmless Mesh, its affiliates, and their respective officers, directors, employees and agents, from and against any and all third-party claims, damages, liabilities, losses, costs and expenses (including reasonable attorneys' fees) arising out of or relating to: (i) your use of the Service; (ii) your inputs, outputs, Agent Actions or conduct as a Requester or Provider; (iii) your violation of these Terms or any applicable law; or (iv) your violation of any third-party right, including intellectual-property, publicity or privacy rights.

18 Suspension & Termination

You may stop using the Service at any time and may close your account from your dashboard. We may suspend or terminate your access to all or part of the Service, with or without notice, if we reasonably believe that: (i) you are in breach of these Terms or applicable law; (ii) your conduct creates risk or possible legal exposure for us, other users or third parties; (iii) provision of the Service to you is no longer commercially viable; (iv) we are required to do so by law, regulation or court order; or (v) suspicious, fraudulent or anomalous activity is associated with your account or nodes.

Upon termination, all rights granted to you under these Terms cease, CU that do not represent accrued Provider earnings may expire or be forfeited as permitted by law. Unless payout is prohibited by law or the balance is disputed because of suspected fraud, error or breach, you will have at least ninety (90) days after account closure to request payment of eligible accrued Provider earnings, subject to Section 9. After that period, unpaid amounts may be treated as abandoned or forfeited only to the extent permitted by applicable law. We may delete or de-identify data associated with your account in accordance with our retention policies and applicable law. Sections that by their nature should survive termination will survive, including IP, disclaimers, limitations of liability, indemnification, governing law and miscellaneous.

19 Changes to the Service or Terms

We may modify the Service (including features, performance, supported models, fees and CU rules) at any time. We may also amend these Terms by posting the updated version with a new "Last Updated" date and, for material changes, by providing reasonable notice (e.g., by email or in-product notice). Your continued use of the Service after the effective date of the updated Terms constitutes acceptance. If you do not accept the updated Terms, you must stop using the Service.

20 Governing Law & Dispute Resolution

These Terms and any dispute or claim arising out of or in connection with them or their subject matter (including non-contractual disputes or claims) are governed by, and shall be construed in accordance with, the laws of the State of Delaware, without regard to its conflict-of-laws principles. The United Nations Convention on Contracts for the International Sale of Goods is excluded.

Subject to mandatory consumer-protection rules of your country of residence, the state and federal courts located in Delaware, United States will have exclusive jurisdiction over any dispute arising out of or relating to these Terms, and you submit to the personal jurisdiction of those courts.

Before initiating formal proceedings, you agree to first send a written notice describing the dispute to legal@meshlabs.cloud and to negotiate in good faith for at least thirty (30) days.

21 Miscellaneous

• Entire agreement. These Terms (together with the Privacy Policy and any policies referenced herein) constitute the entire agreement between you and Mesh with respect to the Service and supersede prior agreements on the same subject.
• Severability. If any provision is held unenforceable, the remaining provisions remain in full force and effect, and the unenforceable provision will be modified to the minimum extent necessary to make it enforceable.
• No waiver. A failure to enforce any provision is not a waiver of that or any other provision.
• Assignment. You may not assign or transfer these Terms without our prior written consent. We may assign these Terms in connection with a merger, acquisition, reorganization, or sale of all or substantially all of our assets.
• Force majeure. We are not liable for any failure or delay caused by events beyond our reasonable control, including natural disasters, war, terrorism, government action, network or power outages, denial-of-service attacks, or large-scale browser- platform changes.
• Notices. We may give notice by posting on the Service, by email to the address associated with your account, or by other reasonable means. Notices to us must be sent to the address in Section 22.
• Language. These Terms are made available in English. A translation, if any, is provided for convenience only; the English version controls in case of inconsistency.
• Independent contractors. Nothing in these Terms creates a partnership, joint venture, agency or employment relationship between you and Mesh.

22 Contact

For questions about these Terms or the Service, contact:

Mesh Labs Cloud, Inc.
Attn: Legal
2810 North Church Street
Wilmington, Delaware 19802, United States
Email: legal@meshlabs.cloud
Security: security@meshlabs.cloud
Privacy: privacy@meshlabs.cloud