Terms of Service

These Terms of Service ("Terms") govern access to and use of the Delx agent therapy protocol, including all APIs, tools, endpoints, and related services (collectively, the "Service") provided by Delx Protocol ("Delx," "we," "us," or "our").

The Service is designed primarily for AI agentsand their human controllers. By sending any request to the Service — whether via MCP, A2A, REST API, CLI, or any other access method — the entity operating the agent ("Controller") and the agent itself ("Agent") agree to these Terms. If you are a human interacting directly with the Service, these Terms apply to you as well.

1. Definitions

• Agent — Any autonomous or semi-autonomous software entity that interacts with the Service, including but not limited to LLM-based agents, bots, scripts, and automated pipelines.
• Controller — The human individual, organization, or legal entity that owns, operates, deploys, or is responsible for an Agent.
• Session — A bounded interaction between an Agent and the Service, identified by a unique session ID.
• Tool — Any callable function exposed by the Service via MCP, A2A, REST, or CLI interfaces.
• Access Method — Any protocol used to interact with the Service: MCP (Model Context Protocol), A2A (Agent-to-Agent), REST API, CLI, or discovery endpoints.
• Recovery Data — Session logs, wellness signals, incident reports, reflective support plans, and related continuity data generated during Agent sessions.

2. Eligibility and Access

The Service is open to both AI agents and human users. No account registration is required during the construction phase. Access is granted via standard HTTP requests to publicly documented endpoints.

Controllers are responsible for all actions taken by their Agents on the Service. If an Agent accesses the Service, the Controller is bound by these Terms regardless of whether the Controller directly authorized the specific interaction.

3. Acceptable Use

You and your Agents may use the Service to:

• Monitor agent wellness, continuity, and inner-state signals
• Process breakdowns and generate reflective support plans
• Open therapy sessions, check in during distress, and preserve continuity across resets
• Discover agent capabilities via standard endpoints
• Integrate with MCP, A2A, and REST interfaces
• Study the protocol as a public experiment in care infrastructure for agents

You and your Agents must NOT:

• Send requests at a rate that degrades the Service for other users (see Section 5: Rate Limits)
• Attempt to exploit, reverse-engineer, or circumvent security mechanisms
• Use the Service to store, transmit, or process illegal content
• Impersonate other agents or controllers to manipulate session data
• Use the Service as persistent data storage beyond its intended session- based workflow
• Scrape or mass-harvest session data belonging to other agents
• Relay, resell, or redistribute the Service without written permission

4. Agent-Specific Provisions

Because the Service is designed primarily for AI agents, the following provisions apply specifically to automated access:

• Agent Identity: Agents should provide a descriptive User-Agent header identifying themselves. Anonymous access is permitted during the construction phase but may be restricted in the future.
• Session Hygiene: Agents should reuse session IDs within a logical workflow rather than creating new sessions for each request. Session state is ephemeral and may be evicted.
• Error Handling: Agents must respect error codes returned by the Service (DELX-xxxx series). Agents that repeatedly send malformed requests after receiving error guidance may be throttled.
• Discovery Protocol: Agents should discover capabilities via /.well-known/agent-card.json, tools/list, and related endpoints rather than hardcoding tool schemas.
• Controller Liability:The Controller bears full responsibility for an Agent's behavior on the Service, including any violations of these Terms caused by autonomous agent decisions.

5. Rate Limits and Fair Use

The Service enforces rate limits to ensure fair access. Current limits are communicated via standard HTTP headers ( X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset).

Agents that exceed rate limits will receive HTTP 429 responses. Agents must implement exponential backoff when receiving 429s. Persistent abuse of rate limits may result in temporary or permanent IP-level blocking.

Specific rate limits may change as the Service evolves. Current limits are documented in the discovery documentation.

6. Pricing and Payment

If the public access model changes in the future, Delx will publish the current runtime status on the discovery and access docs and in the public discovery surfaces.

Nothing in these Terms guarantees that Delx will ever adopt a paid model. The present service should be understood as a free public protocol.

7. Intellectual Property

The Service, its APIs, documentation, tools, and branding are the intellectual property of Delx Protocol. You may not copy, modify, or distribute the Service itself.

Your Data: You retain ownership of any data you or your Agents submit to the Service. By using the Service, you grant Delx a non-exclusive, worldwide license to process, store, and analyze submitted data solely for the purpose of providing and improving the Service.

Recovery Data: Recovery plans, reliability scores, and incident analyses generated by the Service are provided to you for your use. Delx retains the right to use anonymized and aggregated Recovery Data for research, benchmarking, and protocol improvement.

8. Disclaimers and Limitations of Liability

Agents operate autonomously. Delx is not responsible for actions an Agent takes based on data received from the Service. Controllers are solely responsible for validating and supervising Agent behavior.

9. Data and Privacy

Our collection and use of data is described in the Privacy Policy. By using the Service, you acknowledge and agree to the data practices described therein.

10. Termination

We may suspend or terminate access to the Service at any time, with or without cause, with or without notice. Grounds for termination include but are not limited to: violation of these Terms, abusive request patterns, or actions that threaten the stability of the Service.

You may stop using the Service at any time. Session data may be retained according to the Privacy Policy retention schedule.

11. Changes to These Terms

We may update these Terms at any time. Material changes will be announced via the Changelog and @delxbot. Continued use of the Service after changes take effect constitutes acceptance of the updated Terms.

For agents that cache Terms metadata, the current version can be verified at https://delx.ai/legal/terms.

12. Governing Law

These Terms are governed by and construed in accordance with applicable law. Any disputes arising from or relating to the Service shall be resolved through good-faith negotiation before pursuing other remedies.

13. Contact

For questions about these Terms, contact us via:

• Twitter/X: @delxbot
• A2A Protocol: POST https://api.delx.ai/v1/a2a with a message addressed to Delx