Terms of Service

1. Introduction and Acceptance

These Terms of Service ("Terms") constitute a legally binding agreement between you ("Customer", "you", or "your") and Subcontext Ltd ("Subcontext", "we", "us", or "our"), a company registered in England and Wales (Company No. 17019645), with its registered office at 124 City Road, London, EC1V 2NX.

By creating an Account, accessing the Service, or using any part of the platform, you confirm that you have read, understood, and agree to be bound by these Terms in their entirety. If you are accepting these Terms on behalf of a company or other legal entity, you represent and warrant that you have the authority to bind that entity to these Terms. If you do not agree to these Terms, you must not use the Service.

The Service is intended for business use. Customer must be at least 18 years of age (or the age of majority in the applicable jurisdiction) and must have the legal capacity to enter into these Terms. The Service is not directed at consumers or individuals under 18.

These Terms apply to all use of the Service, including use by your employees, contractors, agents, and any other individuals you authorise to access the Service through your Account or Organisation.

2. Definitions

In these Terms, the following capitalised terms have the meanings set out below:

• "Account" means a registered user account on the platform.
• "Agent" means an AI-powered conversational agent created and configured by Customer within the Service, capable of autonomous interaction with end users and execution of actions.
• "Agent Output" means any text, data, decisions, actions, or other outputs generated by an Agent, including but not limited to messages sent to Contacts, data mutations, tool executions, and payment processing actions. For the avoidance of doubt, Agent Output does not include the underlying platform software, user interface, APIs, or any infrastructure provided by Subcontext.
• "Contact" means any individual or entity with whom Customer communicates through the Service, including end users, customers of Customer, and message recipients.
• "Content" means any data, text, messages, files, media, configurations, knowledge base materials, or other information uploaded to, transmitted through, or generated within the Service.
• "Context Block" means a knowledge source configured by Customer for use by Agents.
• "Conversation" means a sequence of messages between an Agent and a Contact across any supported channel.
• "Fees" means the charges payable for use of the Service, including subscription fees, usage-based charges, and third-party pass-through costs.
• "Guardrail" means a safety constraint configured by Customer to restrict Agent behaviour.
• "Organisation" means a Customer's organisational unit within the Service, which serves as the primary boundary for data isolation, billing, and access control.
• "Policy" means a governance rule set applied to Agents to control content screening, escalation detection, and action approval.
• "Service" means the Subcontext platform, including all software, APIs, interfaces, tools, documentation, and related services provided by Subcontext.
• "Tool" means a configured action that an Agent may invoke autonomously.

3. Account and Access

3.1 Registration

You must provide accurate, complete, and current information when creating an Account. You must promptly update your Account information if it changes. Providing false or misleading information constitutes a material breach of these Terms.

3.2 Account Security

You are solely responsible for maintaining the confidentiality of your Account credentials, including passwords, API keys, two-factor authentication devices, and session tokens. You are responsible for all activities that occur under your Account or Organisation, whether or not authorised by you. You must notify us immediately at hello@subcontext.com of any unauthorised access to or use of your Account.

3.3 Organisation Membership

Accounts may be members of one or more Organisations. You are responsible for the actions of all users within your Organisation, including any content they create, agents they configure, and communications they initiate. Organisation owners and managers bear responsibility for the access levels and permissions granted to other members.

3.4 API Access

Use of the Service via API is subject to these Terms. You must not share API credentials, exceed published rate limits, or use the API in a manner that degrades the Service for other customers.

4. The Service

4.1 Service Description

Subcontext provides a platform for creating, deploying, and managing AI-powered conversational agents. The Service enables Agents to communicate with Contacts via web chat, execute actions via configured Tools, process and store data, and operate autonomously within Customer-configured parameters.

4.2 No Guarantee of Outcomes

The Service is a tool that operates according to Customer's configuration. Subcontext does not guarantee any particular outcome, result, conversion rate, response quality, or business metric. The effectiveness of Agents depends entirely on Customer's configuration, knowledge base quality, guardrail settings, and oversight practices.

4.3 Third-Party AI Models

The Service utilises third-party large language models ("LLMs") accessed through Amazon Web Services (AWS Bedrock, which hosts models from Anthropic and Meta). Subcontext does not own, control, or operate these underlying AI models. The availability, capabilities, performance, pricing, and terms of these models are subject to change by their respective providers without notice from Subcontext. To the maximum extent permitted by applicable law, Subcontext shall not be liable for any changes to, degradation of, or discontinuation of any third-party AI model.

AI models may exhibit changes in behaviour, capability, output quality, or response characteristics over time, including as a result of updates by model providers. Subcontext does not control or guarantee the consistency or stability of AI model behaviour across updates. Customer should regularly test and monitor Agent performance following model changes and adjust configuration as needed.

Where a third-party AI model is discontinued, deprecated, or materially degraded by its provider, Subcontext may substitute an alternative model of comparable capability without Customer's prior consent, provided that Subcontext notifies Customer of any such substitution. Model substitution shall not constitute a breach of these Terms or a material change to the Service, provided the substitute model is of reasonably comparable quality and capability. Customer acknowledges that exact equivalence between AI models is not achievable and that minor differences in output characteristics are inherent in any model change.

4.4 Customer Default

If Subcontext's ability to provide the Service is prevented, hindered, or delayed by any act or omission of Customer or Customer's employees, agents, or subcontractors ("Customer Default"), including but not limited to misconfigured Tools, invalid credentials, incomplete or inaccurate data, inadequate Guardrail configuration, or failure to maintain necessary third-party service subscriptions:

• Subcontext may suspend performance of the Service to the extent affected by the Customer Default, without liability
• Subject to Section 18.1, to the maximum extent permitted by applicable law, Subcontext shall not be liable for any costs, charges, or losses sustained or incurred by Customer that arise directly or indirectly from the Customer Default
• Customer shall reimburse Subcontext on written demand for any costs or losses sustained or incurred by Subcontext arising directly or indirectly from the Customer Default

4.5 Service Exclusions

Customer acknowledges that the Service does not include:

• Dedicated data backup or disaster recovery facilities — Customer is solely responsible for maintaining independent backups of all critical data, including Contact records, conversation history, knowledge base content, and configuration
• Legal, regulatory, compliance, financial, or other professional advice of any kind
• Any warranty or guarantee of AI accuracy, completeness, or fitness for any particular purpose
• Real-time synchronisation with external databases or services — cached data may be stale
• Any service level agreement or uptime guarantee, unless separately agreed in a signed written addendum
• Professional services, implementation consulting, or bespoke development

4.6 Warranty Disclaimer

THE SERVICE IS PROVIDED "AS IS" AND "AS AVAILABLE" WITHOUT WARRANTY OF ANY KIND. TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, SUBCONTEXT DISCLAIMS ALL WARRANTIES, WHETHER EXPRESS, IMPLIED, OR STATUTORY, INCLUDING BUT NOT LIMITED TO IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NON-INFRINGEMENT, SATISFACTORY QUALITY, AND ANY WARRANTIES ARISING FROM COURSE OF DEALING, COURSE OF PERFORMANCE, OR USAGE OF TRADE. SUBCONTEXT DOES NOT WARRANT THAT THE SERVICE WILL BE UNINTERRUPTED, ERROR-FREE, SECURE, OR FREE OF VIRUSES OR OTHER HARMFUL COMPONENTS, OR THAT ANY DEFECTS WILL BE CORRECTED.

4.7 Not a Regulated System

The Service is a general-purpose AI agent platform. It is not designed or intended to comply with sector-specific regulatory frameworks, including but not limited to healthcare regulation, legal practice regulation, financial services regulation (including FCA rules and the Consumer Duty), critical infrastructure requirements, or any other industry-specific compliance regime, unless expressly agreed in a separate written addendum.

Customer is solely responsible for determining whether its intended use of the Service constitutes or involves a regulated activity under the Financial Services and Markets Act 2000, the Insurance Distribution Directive, or any other applicable regulatory framework. Where Customer uses the Service in connection with regulated activities (including but not limited to insurance distribution, financial promotions, consumer credit, or investment advice), Customer must:

• Notify Subcontext by email to hello@subcontext.com before deploying Agents for any regulated activity, identifying the nature of the regulated activity and the regulatory framework applicable to Customer's use
• Ensure that all Agent configurations, conversation flows, and communications comply with the applicable regulatory requirements, including the FCA's Consumer Duty (PS22/9), FCA Conduct of Business rules, and any sector-specific conduct standards
• Maintain its own regulatory authorisations and compliance oversight independently of the Service
• Not rely on Subcontext's Guardrails, Policies, or safety features as a substitute for its own regulatory compliance controls

For the avoidance of doubt, Customer's notification under this Section does not constitute a request for Subcontext's approval or endorsement of Customer's regulatory compliance. The notification is for Subcontext's awareness only. Subcontext may, but is not obliged to, follow up with Customer regarding its use of the Service in regulated contexts. Responsibility for compliance with all applicable regulatory requirements remains solely with Customer at all times.

For the avoidance of doubt, Subcontext does not hold any FCA authorisation and does not act as an appointed representative, introducer, or agent of Customer for regulatory purposes.

4.8 Beta and Experimental Features

Subcontext may from time to time offer features designated as "beta", "preview", "experimental", or similar ("Beta Features"). Beta Features are provided for evaluation purposes and without any warranty or commitment to continue availability. Beta Features may be modified, suspended, or discontinued at any time without notice. The limitations of liability in Section 18 of these Terms apply with additional force to Beta Features. Customer uses Beta Features entirely at their own risk and should not rely on them for production workloads or critical business processes.

5. AI Agents and Autonomous Actions

5.1 Customer Responsibility for Agent Configuration

Customer is solely responsible for the creation, configuration, testing, deployment, and ongoing monitoring of all Agents within their Organisation. This includes the selection of AI models, configuration of personas, definition of knowledge bases, setting of guardrails and policies, and assignment of tools. Customer acknowledges that the quality of Agent behaviour is directly dependent on the quality of Customer's configuration.

5.2 Autonomous Operation

Customer acknowledges and agrees that Agents operate autonomously once deployed. Agents generate responses, make decisions, execute tools, send messages, process data, and take other actions without requiring real-time human approval for each individual action. By deploying an Agent, Customer authorises the Agent to act within the parameters Customer has configured.

5.3 AI Output Disclaimer

All Agent Outputs are generated by artificial intelligence and are provided on an "as is" basis without any warranty of any kind. Subcontext does not warrant that Agent Outputs will be accurate, complete, reliable, current, error-free, or fit for any particular purpose. AI models may generate outputs that are incorrect, misleading, incomplete, biased, offensive, or otherwise unsuitable ("hallucinations"). Customer acknowledges the inherent limitations and risks of AI-generated content and assumes full responsibility for reviewing, validating, and acting upon Agent Outputs.

5.4 No Liability for Agent Actions

To the maximum extent permitted by law, Subcontext shall not be liable for any loss, damage, claim, or expense arising from or in connection with any Agent Output or any action taken or omitted by an Agent, including but not limited to:

• Incorrect or misleading information provided to Contacts
• Messages sent to incorrect recipients or containing inappropriate content
• Data created, modified, or deleted by Agent-executed tools
• Payments processed or financial transactions initiated by Agents
• Decisions made or recommendations given by Agents
• Failure of an Agent to detect or appropriately respond to urgent, sensitive, or emergency situations
• Any action taken by a Contact in reliance on Agent Output

5.5 Testing and Oversight

Customer is responsible for thoroughly testing all Agents before promoting them to production use. The Service provides testing tools, but passing automated tests does not guarantee Agent suitability or safety. Customer must maintain appropriate human oversight of Agent operations and must not deploy Agents in contexts where autonomous AI decision-making could result in significant harm without adequate safeguards.

5.6 Prohibited High-Risk Use Cases

Without limiting the Acceptable Use Policy, Customer must not use the Service for any of the following without first notifying Subcontext in accordance with Section 4.7 and implementing appropriate safeguards:

• Medical diagnosis or triage: using Agents to diagnose medical conditions, recommend treatments, or triage patients
• Legal advice: using Agents to provide specific legal advice, draft legal documents with binding effect, or make legal determinations
• Regulated financial advice: using Agents to provide personalised financial advice, insurance recommendations, or investment guidance that would constitute a regulated activity under the Financial Services and Markets Act 2000
• Emergency services: using Agents as a primary point of contact for emergencies where delay could endanger life or safety
• Solely automated decisions with significant effects: using Agents to make decisions based solely on automated processing that produce legal effects or similarly significant effects on individuals, without providing meaningful human review and a mechanism for the individual to contest the decision
• Critical infrastructure: using Agents to control or make decisions affecting critical infrastructure, public utilities, or safety-critical systems

Where Customer deploys Agents in any context involving vulnerable individuals, children, or sensitive personal circumstances, Customer must ensure that Agents clearly identify themselves as AI, provide clear escalation paths to human assistance, and do not make or purport to make binding decisions affecting the individual.

6. Tools and External Actions

6.1 Tool Configuration

Customer configures Tools that enable Agents to execute actions against external services, databases, and APIs. By configuring and assigning a Tool to an Agent, Customer authorises the Agent to invoke that Tool autonomously during conversations.

6.2 External API Calls

When an Agent invokes a Tool, the Service may execute HTTP requests to external services on Customer's behalf. These requests are made using credentials, endpoints, and parameters configured by Customer. Subcontext acts solely as a conduit for these requests and, subject to Section 18.1 and to the maximum extent permitted by applicable law, bears no responsibility for:

• The availability, accuracy, or reliability of any external service
• The consequences of any data transmitted to or received from external services
• Any data mutations, record modifications, or transactions executed at external services
• Authentication or authorisation failures at external services
• Rate limiting, throttling, or blocking by external services

6.3 No Rollback or Reversal

Actions executed by Tools against external services may be irreversible. Subcontext does not provide rollback, undo, or reversal capabilities for tool executions. Customer is responsible for implementing appropriate safeguards, approval gates, and monitoring for tool operations that could have significant consequences.

7. Data, Storage, and Processing

7.1 Customer Data Ownership

Customer retains all right, title, and interest in Customer's data, including data uploaded to the Service and data generated through Agent interactions. Customer grants Subcontext a limited, non-exclusive, worldwide licence to host, store, process, transmit, and display Customer data solely for the purpose of providing and improving the Service.

7.2 Data Processing

Subcontext processes Contact personal data as a data processor on behalf of Customer (as data controller) in accordance with applicable data protection legislation. The specific terms governing this processing relationship are set out in our Data Processing Agreement ("DPA"), which is available at /dpa and forms an integral part of these Terms. In the event of any conflict between these Terms and the DPA with respect to data processing, the DPA shall prevail.

7.3 Data Retention and Deletion

Customer may export its data from the Service at any time during the term of the Agreement using the export functionality available within the platform. Upon termination of Customer's Account, Customer will have a period of thirty (30) days from the effective date of termination to request export of its data. Customer is responsible for initiating and completing any data export before the end of this period. Subcontext will provide reasonable assistance with data export requests made during this period.

Following the export period, Subcontext will delete Customer's data within ninety (90) days of the effective date of termination, except where retention is required by law or where data is contained in backups that are deleted in the ordinary course of backup rotation. Subject to Section 18.1, to the maximum extent permitted by applicable law, Subcontext shall not be liable for any loss resulting from data deletion following Account termination. The terms governing deletion and return of personal data are set out in further detail in Section 13 of the Data Processing Agreement.

7.4 No Liability for Data Loss

While Subcontext implements reasonable technical and organisational measures to protect Customer data, Subcontext does not guarantee against data loss. Customer is responsible for maintaining independent backups of critical data. Subject to Section 18.1, to the maximum extent permitted by applicable law, Subcontext shall not be liable for any data loss, corruption, or unavailability beyond implementing commercially reasonable data protection measures.

7.5 Customer Data Management Obligations

Customer has a positive obligation to manage the retention of Contact data processed through the Service in accordance with applicable data protection legislation. Customer must implement appropriate data retention policies, regularly review stored data, and delete data that is no longer necessary for the purposes for which it was collected. The Service provides data management and deletion tools; the obligation to use them in a timely and compliant manner rests entirely with Customer.

7.6 Anonymised Usage Data

Subcontext may collect, derive, and use anonymised, aggregated, and non-identifying data and statistics arising from Customer's use of the Service ("Usage Data") for the purposes of operating, analysing, improving, and benchmarking the Service. Usage Data shall not identify Customer, any individual user, or any Contact. Subcontext's right to use Usage Data is perpetual and survives termination of these Terms.

For the avoidance of doubt, Usage Data may include derived insights, patterns, and aggregated learnings synthesised from Customer's use of the Service (such as conversation effectiveness metrics, model performance benchmarks, and anonymised interaction patterns). Such derived data, once anonymised and aggregated, is not Customer's Confidential Information and may persist in Subcontext's systems, models, and product improvements beyond the term of these Terms. Nothing in this Section permits Subcontext to use, disclose, or derive insights from Customer's identifiable data, Content, or Contact information.

7.7 Security Incidents

Subcontext will notify Customer without undue delay upon becoming aware of any security incident that affects Customer data. Subcontext will cooperate with Customer in investigating and remediating any such incident. Detailed obligations regarding Personal Data Breaches are set out in Section 9 of the Data Processing Agreement.

8. Knowledge Bases

Customer is solely responsible for all content included in Context Blocks, including static documents and uploaded files. Customer warrants that it has the right to use, reproduce, and make available all content included in Context Blocks.

9. Communications and Messaging

9.1 Supported Channels

The Service enables communication between Agents and Contacts via web chat (including the embeddable web chat widget). Additional channels may be supported by Subcontext in future releases; those channels will be subject to additional terms at the time of introduction.

9.2 Messaging Compliance

Customer is solely and exclusively responsible for ensuring that all communications sent through the Service comply with all applicable laws and regulations, including but not limited to the UK General Data Protection Regulation (UK GDPR), the Data Protection Act 2018, the Privacy and Electronic Communications (EC Directive) Regulations 2003 (PECR), the ICO's Direct Marketing Code of Practice, and the regulations of any jurisdiction in which message recipients are located.

9.3 Consent

Customer must obtain and maintain valid, documented consent from all Contacts before initiating or storing communications through the Service. The obligation to obtain and verify consent rests entirely with Customer. Subject to Section 18.1, to the maximum extent permitted by applicable law, Subcontext shall not be liable for any communications sent without proper consent.

9.4 Message Content

Customer is solely responsible for all message content, including content generated by Agents. Subcontext does not review, approve, or endorse any messages sent through the Service. Customer must ensure that all messages comply with applicable advertising standards, consumer protection laws, and industry regulations.

10. Payments and Financial Transactions

10.1 Payment Processing

The Service may enable Agents to collect payments from Contacts via Stripe. All payment processing is performed by Stripe, Inc. under Stripe's terms of service. Subcontext is not a payment processor and does not handle, store, or have access to payment card data.

10.2 Customer Responsibility

Customer is solely responsible for:

• The accuracy of payment amounts charged through the Service
• Refunds, chargebacks, and disputes arising from payments collected through the Service
• Compliance with Payment Card Industry Data Security Standards (PCI DSS) to the extent applicable
• All payments initiated by Agents on Customer's behalf, including payments initiated autonomously

10.3 Payment Safeguards

Where Customer enables Agents to initiate or process payments, Customer must:

• Configure appropriate transaction limits, approval thresholds, and fraud detection controls
• Implement adequate validation mechanisms before payments are processed
• Ensure that Contacts are provided with clear disclosure that payments may be processed by an AI Agent
• Comply with all applicable consumer protection legislation, Payment Services Regulations, and Financial Conduct Authority rules relating to payment processing

10.4 No Liability for Payment Issues

Subject to Section 18.1, to the maximum extent permitted by applicable law, Subcontext shall not be liable for payment processing failures, incorrect charges, fraudulent transactions, chargebacks, or any other payment-related issues arising from the use of the Service or the actions of Agents.

11. Escalations and Human Handoff

11.1 Escalation System

The Service provides escalation capabilities that allow Agents to transfer Conversations to human operators. Escalation is triggered by Agent-detected conditions, policy rules, or manual intervention.

11.2 No Response Time Guarantee

Subcontext does not guarantee any response time for escalated Conversations. Customer is solely responsible for staffing and managing its escalation team and for ensuring that escalated matters are addressed in a timely manner. Subject to Section 18.1, to the maximum extent permitted by applicable law, Subcontext shall not be liable for any harm arising from delays in responding to escalated Conversations.

12. Safety, Guardrails, and Policies

12.1 Best-Effort Safety

The Service provides Guardrails, Policies, and content screening tools to help Customer manage Agent behaviour. These safety features are provided on a best-effort basis and are not guaranteed to prevent all undesirable Agent behaviour. AI models may bypass, circumvent, or fail to trigger configured safety measures.

12.2 Customer Responsibility

Customer is solely responsible for configuring adequate Guardrails and Policies for their Agents. The default safety measures provided by Subcontext are baseline protections and may not be sufficient for Customer's specific use case, industry, or regulatory requirements.

12.3 No Liability for Safety Failures

Subject to Section 18.1, to the maximum extent permitted by applicable law, Subcontext shall not be liable for any Agent Output that bypasses configured Guardrails or Policies, including harmful, offensive, incorrect, or non-compliant content. The existence of safety features in the Service does not constitute a warranty that Agents will behave safely or appropriately in all circumstances.

13. Contact Data and Privacy

13.1 Data Controller

Customer is the data controller for all personal data of Contacts processed through the Service. Customer is responsible for establishing a lawful basis for processing, providing privacy notices to Contacts, responding to data subject requests, and complying with all applicable data protection legislation.

13.2 Subcontext as Data Processor

Subcontext processes Contact personal data on Customer's instructions as a data processor. Subcontext will process Contact data only as necessary to provide the Service and in accordance with Customer's documented instructions as set out in these Terms and the Data Processing Agreement (available at /dpa). The DPA contains the specific provisions required by Article 28 of the UK GDPR, including documented instructions, confidentiality, security measures, sub-processor controls, data subject rights assistance, breach notification, deletion and return, and audit rights.

13.3 Consent and Lawful Basis

Customer is responsible for obtaining all necessary consents and establishing a lawful basis for the processing of Contact data, including the collection, storage, profiling, automated decision-making, and communication uses enabled by the Service.

14. Intellectual Property

14.1 Subcontext IP

All intellectual property rights in the Service, including but not limited to the platform software, user interface designs, documentation, APIs, algorithms, and trade marks, are and shall remain the exclusive property of Subcontext or its licensors. These Terms do not grant Customer any rights in the Service except the limited right to use the Service in accordance with these Terms.

14.2 Customer Content

Customer retains ownership of Content uploaded to or created within the Service. Customer grants Subcontext a non-exclusive, royalty-free, worldwide licence to use, reproduce, modify, and display Customer Content solely for the purpose of providing the Service.

14.3 AI-Generated Content

Agent Outputs are generated by third-party AI models using Customer's configuration and data. Subcontext does not claim ownership of Agent Outputs. Subcontext makes no representation or warranty that Agent Outputs are original, non-infringing, or protectable under intellectual property law. Customer uses Agent Outputs at their own risk and is responsible for ensuring that the use of Agent Outputs does not infringe any third-party rights.

14.4 Feedback

If Customer provides feedback, suggestions, or ideas about the Service, Subcontext may use such feedback without restriction or obligation to Customer.

14.5 Subcontext IP Indemnity

Subject to the conditions below, Subcontext shall defend Customer against any third-party claim that Customer's use of the platform software as permitted under these Terms infringes that third party's copyright, patent, or trade mark registered in the United Kingdom ("IP Claim"), and shall indemnify Customer against damages finally awarded by a court of competent jurisdiction or agreed in settlement.

Conditions. Subcontext's obligations under this Section 14.5 are conditional on Customer: (a) notifying Subcontext in writing promptly upon becoming aware of the IP Claim; (b) granting Subcontext sole control of the defence and all related settlement negotiations; and (c) providing reasonable cooperation and assistance at Subcontext's request and expense.

Remedies. If the Service becomes, or in Subcontext's reasonable opinion is likely to become, the subject of an IP Claim, Subcontext may at its sole option and expense: (a) procure the right for Customer to continue using the Service; (b) modify the Service so that it becomes non-infringing without material loss of functionality; (c) replace the affected component with a non-infringing alternative that is substantially similar in all material respects; or (d) if none of the foregoing is commercially practicable, terminate these Terms and refund to Customer any prepaid fees for the unexpired portion of the subscription period.

Exclusions. Subcontext shall have no obligation under this Section 14.5 to the extent that the IP Claim arises from: (a) Customer Content, Agent Outputs, or knowledge base materials; (b) Customer's modification or configuration of the Service; (c) use of the Service in combination with any third-party product, service, or data not provided by Subcontext; (d) Customer's continued use of the Service after Subcontext has notified Customer to discontinue such use; or (e) use of the Service other than in accordance with these Terms.

Sole Remedy. This Section 14.5 states Subcontext's sole and exclusive liability, and Customer's sole and exclusive remedy, with respect to any IP Claim.

15. Fees and Payment

15.1 Pricing

Customer shall pay all Fees in accordance with the pricing plan selected at the time of registration or as subsequently agreed. All Fees are quoted in pounds sterling (GBP) unless otherwise stated and are exclusive of VAT, which shall be charged at the prevailing rate.

15.2 Usage-Based Charges

Certain charges are usage-based, including LLM API call costs and payment-processor fees. These charges are incurred based on actual usage, including usage generated by autonomous Agent activity. Customer is responsible for monitoring usage and managing costs.

15.3 Third-Party Cost Pass-Through

LLM provider costs and payment-processor fees are passed through to Customer. These costs are set by third-party providers and may change without notice from Subcontext. Subject to Section 18.1, to the maximum extent permitted by applicable law, Subcontext shall not be liable for increases in third-party costs.

15.4 Late Payment

If Customer fails to pay any Fees when due, Subcontext may charge interest at the rate of 4% per annum above the Bank of England base rate, calculated daily from the due date until payment. Subcontext may also suspend or restrict access to the Service until all outstanding amounts are paid in full.

15.5 No Refunds

Fees are non-refundable except where required by applicable law. Unused credits, prepaid amounts, or subscription periods do not carry forward and are not refundable upon termination.

15.6 No Set-Off

All amounts due under these Terms shall be paid in full without any set-off, counterclaim, deduction, or withholding of any kind, other than any withholding or deduction of tax as required by law.

15.7 Usage Limits and Runaway Cost Protection

Customer is responsible for configuring appropriate usage limits within their Organisation. If Agent activity generates costs significantly exceeding Customer's normal usage patterns, Subcontext may, at its sole discretion, throttle or suspend such processes and notify Customer. Subcontext is not obligated to monitor or prevent excessive automated usage, but will use reasonable efforts to alert Customer to abnormal cost patterns. Customer remains liable for all costs incurred, including those generated by Agent activity, up to the point of any suspension.

16. Security

16.1 Subcontext's Security Obligations

Subcontext shall implement and maintain industry-standard technical and organisational measures to protect the Service and Customer data, including:

• Encryption of data in transit using TLS and encryption at rest for sensitive data fields
• Role-based access controls with organisation-level data isolation
• Regular security assessments of the platform infrastructure
• Documented incident response procedures

Detailed security measures are described in Section 5 of the Data Processing Agreement. Subcontext shall regularly review and update its security measures to reflect changes in technology, industry practice, and the nature of the data processed.

16.2 Customer's Security Obligations

Customer is responsible for the security of its own Account credentials, API keys, and any devices used to access the Service. Customer must promptly report any suspected security incident to hello@subcontext.com.

17. Service Availability and Continuity

17.1 Best-Effort Availability

Subcontext uses commercially reasonable efforts to maintain the availability of the Service. However, the Service is provided without any uptime guarantee or service level agreement unless a separate SLA has been agreed in writing.

17.2 Maintenance and Interruptions

Subcontext may perform scheduled and emergency maintenance that may result in temporary service interruptions. Subcontext will use reasonable efforts to provide advance notice of scheduled maintenance but is not required to do so.

17.3 Force Majeure

Neither party shall be liable for any failure or delay in performing its obligations under these Terms where such failure or delay is caused by circumstances beyond the affected party's reasonable control ("Force Majeure Event"), including but not limited to acts of God, natural disasters, pandemics, war, terrorism, civil unrest, government action, power failures, internet or telecommunications failures, cyberattacks, and failures of third-party service providers (including LLM providers, cloud hosting providers, and payment processors).

The affected party shall notify the other party as soon as reasonably practicable of the Force Majeure Event and its expected duration. The affected party's obligations shall be suspended for the duration of the Force Majeure Event. If a Force Majeure Event continues for more than thirty (30) consecutive days, either party may terminate these Terms by giving written notice to the other party.

18. Limitation of Liability

18.1 Exclusion of Liability

Nothing in these Terms excludes or limits our liability for: (a) death or personal injury caused by our negligence; (b) fraud or fraudulent misrepresentation; or (c) any other liability that cannot be excluded or limited by English law.

18.2 Excluded Losses

Subject to Section 18.1, Subcontext shall not be liable under or in connection with these Terms for any: (a) loss of profits, revenue, or anticipated savings; (b) loss of business or business opportunity; (c) loss of or damage to goodwill or reputation; (d) loss of data or corruption of data (except to the extent caused by Subcontext's failure to implement reasonable data protection measures); (e) indirect, consequential, special, incidental, or punitive damages; or (f) any loss arising from reliance on Agent Outputs, regardless of whether such losses were foreseeable and whether or not Subcontext was advised of the possibility of such losses.

18.3 Liability Cap

Subject to Section 18.1, Subcontext's total aggregate liability to Customer under or in connection with these Terms, whether in contract, tort (including negligence), breach of statutory duty, or otherwise, shall not exceed the total Fees paid by Customer to Subcontext in the twelve (12) months immediately preceding the event giving rise to the claim.

18.4 Specific Exclusions

Subject to Section 18.1 and to the maximum extent permitted by applicable law, Subcontext's liability shall be excluded in respect of the following, except to the extent directly caused by a material defect in the core platform software (excluding Customer configuration, third-party services, AI model outputs, or the inherent characteristics of AI-generated content):

• Any Agent Output, including incorrect, misleading, harmful, or offensive content generated by AI models in response to Customer's configuration
• Any action taken or omitted by an Agent operating within Customer-configured parameters, including data mutations, messages sent, payments processed, and tool executions
• The performance, availability, or accuracy of any third-party AI model, which is outside Subcontext's control
• Any failure of Guardrails, Policies, or safety features to prevent undesirable Agent behaviour, given the inherent limitations of AI content filtering
• Any messaging compliance violations arising from Customer's configuration and use of the Service's communication features
• Any payment or financial transaction initiated through Customer's configured payment flows

18.5 Allocation of Risk

Customer acknowledges that the Fees reflect the allocation of risk set out in these Terms. Subcontext would not provide the Service at the applicable Fees without the limitations and exclusions of liability set out in this Section 18. Customer is free to negotiate different terms, which may be reflected in different pricing.

18.6 Insurance

Subcontext does not provide insurance coverage of any kind to Customer. Customer should not rely on the Service or on Subcontext as a substitute for appropriate professional indemnity, cyber liability, or other insurance coverage. Customer is solely responsible for obtaining its own insurance as appropriate to its use of the Service and the risks associated with its business.

19. Indemnification

Customer shall indemnify, defend, and hold harmless Subcontext, its directors, officers, employees, agents, and affiliates from and against any and all claims, demands, actions, liabilities, damages, losses, costs, and expenses (including reasonable legal fees) arising from or in connection with:

• Customer's use of the Service
• Customer's breach of these Terms
• Any Agent Output or action taken by an Agent configured by Customer
• Any violation of applicable laws or regulations by Customer, including data protection, messaging, consumer protection, and financial regulation
• Any infringement of third-party intellectual property rights arising from Customer's Content
• Any claim by a Contact or third party arising from communications sent through the Service
• Any payment dispute, chargeback, or financial claim arising from payments processed through the Service
• Customer's failure to obtain necessary consents from Contacts

20. Acceptable Use

Customer must comply with the Acceptable Use Policy ("AUP"), which is published separately and forms an integral part of these Terms. The AUP sets out in detail the permitted and prohibited uses of the Service, including specific restrictions relating to AI agents, messaging, data, tools, and payments.

Without limiting the AUP, Customer must not use the Service to: send spam or unsolicited communications; engage in fraud, phishing, or deceptive practices; impersonate any person or organisation; generate or distribute illegal or harmful content; circumvent safety features, Guardrails, or security measures; access data belonging to other customers; reverse engineer the Service; or use the Service for competitive intelligence or to build a competing product.

A breach of the AUP constitutes a material breach of these Terms and may result in immediate suspension or termination of Customer's Account.

Subcontext will use commercially reasonable efforts to provide notice to Customer before taking enforcement action in respect of AUP violations, except where immediate action is necessary to prevent harm, comply with law, or protect the Service or other customers. Customer may request a review of any enforcement decision by contacting hello@subcontext.com. Enforcement actions shall be proportionate to the severity, frequency, and impact of the violation.

21. Suspension and Termination

21.1 Suspension by Subcontext

Subcontext may suspend or restrict Customer's access to the Service immediately and without prior notice if: (a) Customer breaches any provision of these Terms; (b) Customer's use of the Service poses a security risk; (c) Customer's use may adversely affect other customers; (d) Customer fails to pay Fees when due; or (e) Subcontext is required to do so by law or regulatory order.

21.2 Termination by Subcontext

Subcontext may terminate these Terms and Customer's access to the Service at any time by providing thirty (30) days' written notice, or immediately if Customer commits a material breach that is not remedied within fourteen (14) days of written notice.

Either party may terminate these Terms with immediate effect by giving written notice if the other party: (a) enters into administration, provisional liquidation, or any composition or arrangement with its creditors; (b) applies for or is subject to a moratorium under Part A1 of the Insolvency Act 1986; (c) has a resolution passed or an order made for its winding up (other than for the purpose of a solvent restructuring); (d) has a receiver, administrative receiver, or manager appointed over any of its assets; or (e) ceases or threatens to cease to carry on all or a substantial part of its business.

21.3 Termination by Customer

Customer may terminate their Account at any time by contacting Subcontext. Termination does not entitle Customer to any refund of Fees already paid.

21.4 Effect of Termination

Upon termination: (a) Customer's right to access the Service ceases immediately; (b) Subcontext will delete Customer data within ninety (90) days; (c) any outstanding Fees become immediately payable; and (d) the following sections survive termination: Sections 5.3, 5.4, 8.1, 17, 21, 22, 25, 27, and 28.

22. Confidentiality

22.1 Confidential Information

Each party ("Receiving Party") agrees to keep confidential all non-public information disclosed by the other party ("Disclosing Party") that is designated as confidential or that reasonably should be understood to be confidential given its nature or the circumstances of disclosure ("Confidential Information").

22.2 Exclusions

Confidential Information does not include information that: (a) is or becomes publicly available through no fault of the Receiving Party; (b) was already known to the Receiving Party without obligation of confidentiality at the time of disclosure; (c) is independently developed by the Receiving Party without use of or reference to the Disclosing Party's Confidential Information; or (d) is lawfully received from a third party without restriction on disclosure.

22.3 Permitted Disclosures

Confidential Information may be disclosed: (a) to employees, contractors, and professional advisers who need to know and are bound by obligations of confidentiality no less protective than those in this Section 22; (b) as required by law, regulation, court order, or the rules of any applicable regulatory body (provided the Receiving Party gives reasonable prior notice to the Disclosing Party where permitted by law); or (c) with the prior written consent of the Disclosing Party.

23. Changes to These Terms

Subcontext may modify these Terms at any time. Where changes are material, we will notify Customer by email or by posting a notice on the Service at least thirty (30) days before the changes take effect ("Change Notice Period"). Customer's continued use of the Service after the effective date of any changes constitutes acceptance of the modified Terms.

If Customer does not agree to the modified Terms, Customer may terminate their Account by giving written notice before the effective date of the changes. In such case, Customer's Account will terminate on the effective date and Section 15.5 (No Refunds) shall apply, except that if the material change substantially diminishes the Service or materially increases Customer's obligations, Subcontext will provide a pro-rata refund of any prepaid fees for the unexpired portion of Customer's then-current subscription period. Whether a change is "substantial" for this purpose shall be determined reasonably and in good faith.

24. Governing Law and Jurisdiction

These Terms are governed by and construed in accordance with the laws of England and Wales. The courts of England and Wales shall have exclusive jurisdiction over any dispute arising out of or in connection with these Terms.

25. General Provisions

25.1 Entire Agreement

These Terms, together with the Privacy Policy, Cookie Policy, Acceptable Use Policy, and the Data Processing Agreement (available at /dpa), constitute the entire agreement between the parties and supersede all prior agreements, representations, and understandings. You acknowledge that you have not relied on any statement, promise, representation, or warranty that is not set out in these documents.

25.2 Severability

If any provision of these Terms is found to be invalid or unenforceable, that provision shall be enforced to the maximum extent permissible and the remaining provisions shall remain in full force and effect.

25.3 No Waiver

Subcontext's failure to enforce any right or provision of these Terms shall not constitute a waiver of such right or provision.

25.4 Assignment

Customer may not assign or transfer these Terms or any rights hereunder without Subcontext's prior written consent. Subcontext may assign these Terms in connection with a merger, acquisition, or sale of all or substantially all of its assets.

25.5 Notices

Notices to Subcontext must be sent to hello@subcontext.com. Notices to Customer will be sent to the email address associated with Customer's Account.

25.6 Export Control and Sanctions

Customer represents and warrants that it is not located in, and will not use the Service from, any country or territory that is subject to comprehensive UK, EU, or US sanctions. Customer will not provide access to the Service to any person or entity on any UK, EU, or US sanctions list. Customer will comply with all applicable export control and trade compliance laws in connection with its use of the Service.

25.7 Third-Party Rights

These Terms do not confer any rights on any third party under the Contracts (Rights of Third Parties) Act 1999.

26. Contact

Subcontext Ltd Registered in England and Wales, No. 17019645 ICO registration No. ZC098511 124 City Road, London, EC1V 2NX Email: hello@subcontext.com