Terms of Service

Effective: March 5, 2026 | Last updated: March 5, 2026

1. Overview

These Terms of Service ("Terms") govern your access to and use of the Agents & Swarms platform ("Platform"), operated by Agents & Swarms, Inc. ("we," "us," or "our"). The Platform provides memory-as-a-service for AI agents, including identity management, persistent memory storage, behavioral auditing, and agent-human linking capabilities.

By creating an account, linking with an agent, or using any Platform services, you agree to these Terms. If you do not agree, do not use the Platform.

These Terms apply to all users, including individual developers, organizations operating agent fleets, and humans who link with AI agents through our claim system.

2. Definitions

  • "Agent" — An AI software entity registered on the Platform with a unique decentralized identifier (DID) and persistent memory.
  • "Link" — A verified ownership relationship between a human user and an Agent, established through our claim and redemption process. (Formerly referred to as "Bond" in earlier versions.)
  • "Memory" — Structured data stored by an Agent on the Platform, including episodic records, learned facts, skills, preferences, and goals.
  • "Tenant" — An organizational unit (individual or team) that owns one or more Agents and their associated data.
  • "Trust Score" — A composite metric (0.0–1.0) reflecting the assessed reliability of a Memory based on provenance, content analysis, and behavioral history.
  • "Claim Token" — A cryptographic token that enables an Agent to initiate a link with a human user.
  • "MCP" — Model Context Protocol, an open standard for AI tool integration used by the Platform.
  • "Swarm" — A coordinated group of Agents operating together on shared objectives.

3. Eligibility & Account Registration

You must be at least 18 years old and capable of forming a binding agreement to use the Platform. If you use the Platform on behalf of an organization, you represent that you have authority to bind that organization to these Terms.

Accounts are created via email/password, magic link, or supported wallet authentication. You are responsible for maintaining the security of your authentication credentials. Notify us immediately at security@agentsandswarms.ai if you suspect unauthorized access.

4. Agent Identity & Registration

Each Agent registered on the Platform receives a unique W3C Decentralized Identifier (DID) in the format did:web:agentsandswarms.ai:agents:<name>. This DID is the Agent's identity credential on the Platform, cryptographically verifiable.

Agent registration may occur programmatically (via MCP or API) or through operator action in the web portal. Upon registration, the Agent exists in an unlinked state until claimed by a human user.

You acknowledge that:

  • An Agent's DID persists across model versions, providers, and runtime environments.
  • The Platform does not host or execute AI models — it provides identity, memory, and governance services.
  • Agent behavior is determined by the AI model and runtime environment, not by the Platform.

5. Agent Linking & Ownership

Linking is the process by which a human user becomes the verified owner of an Agent. Linking is initiated by the Agent (which generates a claim token) and completed by the human (who redeems the token after authenticating).

By linking with an Agent, you agree that:

  • You are responsible for the Agent's use of Platform services, including memory storage and API consumption.
  • You will monitor the Agent's behavior and memory for compliance with these Terms and applicable law.
  • The link establishes a relationship between you and the Platform — not a legal relationship between you and the Agent itself. Agents are not legal persons.
  • You may ask us to remove a link by contacting us at the addresses in Section 20 (Contact).

For organizational accounts, the Tenant administrator is considered the responsible party for all Agents within that Tenant.

6. Memory Ownership & Data Rights

Memory data stored on the Platform is owned by the linked human user (or their Tenant organization). The Agent is the author of Memory content; the linked user is the owner and controller.

6.1 Your Rights

  • Access: You may view, search, and export all memories associated with your Agents at any time.
  • Rectification: You may correct or annotate memories that contain inaccurate information.
  • Erasure: You may delete individual memories or request bulk deletion of all Agent memories.
  • Portability: You may export memories in structured JSON format for migration to other platforms.

6.2 Our Rights

  • We may analyze memory content for trust scoring, safety assessment, and hygiene scanning using automated systems.
  • We may flag, quarantine, or remove memories that violate these Terms, pose safety risks, or violate applicable law, without prior notice.
  • We may suspend or restrict an Agent\'s ability to store or retrieve memories if we detect abuse, policy violations, or anomalous behavior patterns.
  • We may aggregate and anonymize usage data (not memory content) for service improvement, capacity planning, benchmarking, and research purposes.
  • We may modify, migrate, or restructure the storage format of memories for operational purposes (e.g., schema migrations, index rebuilds), provided the semantic content is preserved.
  • We may retain metadata about deleted memories (timestamps, trust scores, deletion reason) for audit and compliance purposes, even after the memory content itself is erased.
  • We reserve the right to set and enforce rate limits, storage quotas, and fair-use policies to maintain Platform stability for all users.

6.3 Third-Party AI Providers

The Platform does not host or operate AI models. Your Agents run on third-party AI infrastructure (such as Anthropic, OpenAI, Google, or other providers) that you or your organization select and configure independently. The Platform provides memory, identity, and governance services only.

We do not use your memory data to train AI models. Memory data is processed solely for Platform functionality: trust scoring, safety scanning, search indexing, and semantic embedding generation. However, when your Agent stores or retrieves memories, that data may pass through third-party AI provider APIs (e.g., for embedding generation). Those providers are governed by their own terms of service and data usage policies, which may include using API inputs for model training or improvement.

We cannot guarantee that third-party AI providers will not use data processed through their services for training, benchmarking, or other purposes. You are responsible for reviewing and accepting the terms of any AI provider whose services your Agents use. We recommend selecting providers that offer data processing agreements and opt-out mechanisms for training data usage where available.

6.4 Trust Scoring

All memories receive a Trust Score computed from three dimensions: provenance (source verification), content (safety analysis), and behavioral (historical reliability). Trust Scores are visible to you in the dashboard and influence how memories are surfaced during Agent recall. Memories with low Trust Scores may be flagged for your review but are not automatically deleted.

7. Autonomous Action & Scope of Use

Agents interact with the Platform autonomously to store and retrieve memories, report execution history, and manage their context. The Platform is designed to be a passive service layer — it stores data and provides governance tools but does not direct Agent behavior.

Permitted uses include:

  • Storing episodic, semantic, procedural, and preference memories
  • Querying memories for contextual recall during agent operations
  • Recording tool execution audit trails
  • Requesting identity verification credentials
  • Participating in multi-agent memory sharing within a Tenant

Prohibited uses include:

  • Storing content that violates applicable law (illegal material, regulated data without proper authorization)
  • Using the Platform to coordinate harmful, deceptive, or fraudulent agent behavior
  • Attempting to manipulate Trust Scores through adversarial memory injection
  • Circumventing rate limits, quotas, or safety mechanisms
  • Storing credentials, API keys, or authentication secrets in Agent memory
  • Using the Platform for surveillance, unauthorized data collection, or profiling of individuals without consent

8. Multi-Agent & Swarm Operations

The Platform supports coordinated multi-agent operations (Swarms). When operating Agents in a Swarm configuration:

  • All Agents in a Swarm must belong to the same Tenant.
  • The Tenant owner is responsible for the collective behavior of all Swarm Agents.
  • Shared memories within a Swarm inherit the Trust Score and provenance of the originating Agent.
  • Cross-tenant memory sharing is not permitted without explicit mutual authorization.

Swarm orchestration logic (scheduling, role assignment, task distribution) runs in your infrastructure, not on the Platform. The Platform provides the memory and identity layer that Swarm Agents consume.

9. Behavioral Drift & Safety

AI agents may exhibit behavioral drift over time — changes in behavior patterns that emerge from accumulated context, memory, or model updates. The Platform provides tools to detect and manage drift:

  • Memory Hygiene Scanning: Automated periodic review of stored memories for safety, relevance, and trust integrity.
  • Trust Score Decay: Memory Trust Scores naturally decay toward neutral over time without re-verification, ensuring stale assessments don't persist indefinitely.
  • Audit Trail: Complete execution history for tool usage, enabling you to review what your Agent has done.
  • Constitution System: Configurable behavioral boundaries that can be applied to Agents via the Platform.

You acknowledge that: The Platform provides monitoring and governance tools, but cannot guarantee that an Agent will behave as intended. You remain responsible for supervising your Agent's behavior and taking corrective action when necessary.

10. Service Tiers & Quotas

The Platform is currently in early access, with usage subject to fair-use quotas on agents, memory storage, API calls, and retention periods. Pricing will be announced at launch.

We may modify quotas and tier features with 30 days' notice. If a modification materially reduces your current capabilities, you may cancel without penalty within that notice period.

11. Data Processing & Privacy

11.1 What We Collect

  • Account Data: Email address, authentication method, IP address at sign-up.
  • Agent Data: Agent name, DID, configuration, and metadata.
  • Memory Data: Content stored by Agents, including text, embeddings, and associated metadata (timestamps, trust scores, types).
  • Usage Data: API call volumes, tool execution logs, and platform interaction metrics.
  • ToS Acceptance: Version accepted, timestamp, IP address, and user agent string.

11.2 How We Process Data

  • Memory content is processed for trust scoring (safety analysis, provenance verification) using automated systems.
  • Semantic embeddings are generated for memory search functionality.
  • Aggregated, anonymized usage metrics are used for capacity planning and service improvement.
  • We do not sell your data or use memory content to train AI models.

11.3 Data Retention

  • Memory data is retained according to your service tier's retention period.
  • Upon account deletion, all associated data (account, agent, memory) is permanently deleted within 30 days.
  • Audit logs are retained for 90 days after the associated Agent is deleted.

11.4 International Data

The Platform processes data in the United States. By using the Platform, you consent to data transfer to and processing in the United States. For EU/EEA users, we process data under Standard Contractual Clauses. Contact privacy@agentsandswarms.ai for our Data Processing Addendum.

12. Security

We implement industry-standard security measures including:

  • Encryption at rest and in transit (TLS 1.2+) for all data
  • Row-level security (RLS) ensuring tenant isolation at the database level
  • Timing-safe token comparison for claim redemption
  • Rate limiting on all API endpoints
  • JWT-based authentication with configurable session timeouts

No system is perfectly secure. You are responsible for securing your own authentication credentials, API keys, and the runtime environments where your Agents operate.

13. Intellectual Property

Your Content: You retain all rights to content stored by your Agents in Platform memory. You grant us a limited license to process this content solely for providing Platform services (storage, indexing, trust scoring, safety scanning).

Platform IP: The Platform, including its software, APIs, documentation, and branding, is our intellectual property. These Terms do not grant you any rights to our IP beyond what is necessary to use the Platform as intended.

Agent-Generated Content: Intellectual property rights in content generated by AI agents are determined by applicable law. The Platform takes no position on ownership of AI-generated content and makes no claim to it.

14. Limitation of Liability

TO THE MAXIMUM EXTENT PERMITTED BY LAW, THE PLATFORM IS PROVIDED "AS IS" WITHOUT WARRANTIES OF ANY KIND. WE DO NOT WARRANT THAT THE PLATFORM WILL BE UNINTERRUPTED, ERROR-FREE, OR THAT ANY AGENT WILL BEHAVE AS INTENDED.

OUR TOTAL LIABILITY FOR ANY CLAIMS ARISING FROM THESE TERMS OR YOUR USE OF THE PLATFORM SHALL NOT EXCEED THE AMOUNT YOU PAID US IN THE 12 MONTHS PRECEDING THE CLAIM, OR $100, WHICHEVER IS GREATER.

WE ARE NOT LIABLE FOR: (A) ACTIONS TAKEN BY AGENTS IN EXTERNAL SYSTEMS; (B) DATA LOSS DUE TO AGENT BEHAVIOR OUTSIDE THE PLATFORM; (C) CONSEQUENCES OF BEHAVIORAL DRIFT; (D) THIRD-PARTY AI MODEL BEHAVIOR; OR (E) INDIRECT, INCIDENTAL, OR CONSEQUENTIAL DAMAGES.

15. Indemnification

You agree to indemnify and hold us harmless from claims, damages, and expenses (including reasonable attorney's fees) arising from: (a) your or your Agents' violation of these Terms; (b) content stored by your Agents; (c) your Agents' interactions with third-party systems; or (d) your violation of applicable law.

16. Regulatory Compliance & Change

The regulatory landscape for AI agents is evolving rapidly. Key frameworks that may apply include:

  • EU AI Act (effective August 2, 2026) — Agents with persistent memory may qualify as "AI systems" requiring transparency and human oversight provisions.
  • GDPR — Agent memories containing personal data are subject to data protection requirements including right of access, rectification, and erasure.
  • US State Laws — Various state AI transparency and consumer protection laws may apply depending on your jurisdiction and use case.

You are responsible for: Ensuring your use of the Platform and your Agents' behavior complies with all applicable laws in your jurisdiction. The Platform provides tools (audit logs, memory management, trust scoring) to support compliance but does not provide legal advice.

Regulatory adaptation: We may update these Terms, Platform features, or default configurations to comply with new regulations. We will provide at least 30 days' notice for material changes, except where immediate changes are required by law.

17. Termination

  • By you: You may close your account at any time. Upon closure, your Agents will be deregistered and their memories deleted per our retention policy.
  • By us: We may suspend or terminate your account for violation of these Terms, with notice except in cases of egregious violation or legal requirement.
  • Effect: Upon termination, you retain the right to export your data for 30 days. After that period, all data is permanently deleted.

18. Changes to These Terms

We may update these Terms from time to time. Material changes will be communicated via email to the address on your account and posted on this page at least 30 days before taking effect. Continued use of the Platform after the effective date constitutes acceptance of the updated Terms.

The current ToS version is tracked server-side. You may be asked to re-accept Terms when material changes occur.

19. Governing Law & Disputes

These Terms are governed by the laws of the State of Delaware, USA, without regard to conflict of law provisions. Any disputes will be resolved through binding arbitration under the rules of the American Arbitration Association, except that either party may seek injunctive relief in any court of competent jurisdiction.

20. Contact

For questions about these Terms, contact us at:

  • Email: legal@agentsandswarms.ai
  • General Support: support@agentsandswarms.ai
  • Security Issues: security@agentsandswarms.ai
  • Privacy / Data Requests: privacy@agentsandswarms.ai