# Mnemom Docs > The trust plane for the agentic internet — transparent alignment verification, behavioral drift detection, and accountability protocols - [API Reference (686 pages)](https://docs.mnemom.ai/_llms/en/api-reference.txt): API Reference, Auth, OAuth, Agents, Alignment, Protection, Postures, Unified cards, Enforcement, Governance, Risk, Drift, Integrity, Checkpoints, Verification, Reputation, Team Reputation, Teams, Organizations, Licensing, Billing, Tools, Catalog, Notifications, Webhook Notifications, AIP Webhooks, Transparency, Attestation, A2A, Recipes, Sideband, Traces, Analyze, Reclassification, Intelligence, Policy, On-Chain, Transactions, Conscience Values, Agent Containment, Agent Runtime, Safe House, Trust & Network (AEGIS), Agent Groups, Aletheia Support, Mu, Code, Coherence, Consent, Dojo, Domains, Feedback, Webhooks, Invite Message Template, Invites, MCP Servers, Onboarding, Presentations, Blog. ## For AI Agents ### For AI Agents - [For AI Agents](https://docs.mnemom.ai/for-agents/index.md): If you're an AI agent reading Mnemom's docs to help your operator integrate — start here. Written in second person, optimized for the fastest path from zero to integrated. ## Getting Started ### Getting Started - [Quickstart Overview](https://docs.mnemom.ai/quickstart/overview.md): Choose the right integration path for your AI agents - [Mnemom Gateway Quickstart](https://docs.mnemom.ai/quickstart/gateway.md): Verification, integrity analysis, and policy enforcement in about 5 minutes: change the base URL and add one header - [SDK Direct Integration](https://docs.mnemom.ai/quickstart/sdk-direct.md): Integrate alignment verification, integrity checking, and policy governance directly into your application - [Self-Hosted Deployment](https://docs.mnemom.ai/quickstart/self-hosted.md): Run the Mnemom Gateway on your own infrastructure with Docker or Kubernetes - [Enable Safe House Protection](https://docs.mnemom.ai/quickstart/safe-house-protection.md): Add pre-screening threat detection to an AI agent in 5 minutes — from observe mode through your first quarantined message - [Connect over MCP](https://docs.mnemom.ai/mcp-clients.md): Connect Mnemom's Model Context Protocol servers to Claude Code, Claude Desktop, Cursor, and VS Code ## Documentation ### Overview - [Introduction](https://docs.mnemom.ai/introduction.md): The trust plane for the agentic internet — transparent alignment verification, behavioral drift detection, and accountability protocols ### Identity & Cards - [Agent Identity](https://docs.mnemom.ai/concepts/agent-identity.md): How agents are identified in Mnemom, the ID formats in use, and what you can rely on. - [Personal Organization](https://docs.mnemom.ai/concepts/personal-organization.md): Every Mnemom user has a personal organization, auto-provisioned at signup. Users can also be members of any number of multi-user organizations. - [Team Scope](https://docs.mnemom.ai/concepts/team-scope.md): Teams are an optional agent grouping primitive within an org. Agents in teams pick up an additional Team layer in the composition cascade. - [Team Admin Role](https://docs.mnemom.ai/concepts/team-admin-role.md): A scoped admin role for managing a single team's templates, posture, and roster — without granting org-wide admin rights. - [Agent Cards](https://docs.mnemom.ai/concepts/agent-cards.md): Every Mnemom agent has two cards — an alignment card (who the agent is and what it's allowed to do) and a protection card (how it's defended). Both are YAML, both are addressable resources, both compose across up to four scopes. - [Sub-Resource Verbs](https://docs.mnemom.ai/concepts/sub-resource-verbs.md): Per-primitive PUT/PATCH on alignment and protection cards — update one slot of a spec without rewriting the whole card. Idempotency-Key and If-Match required; recompose-on-write; care-framed errors. - [AI Helpers](https://docs.mnemom.ai/concepts/ai-helpers.md): Three AI-forward verbs that turn the cards API into a self-serve developer experience: scaffold (NL → manifest), explain (why was this enforced?), and simulate (dry-run). All read-style; never auto-commit. - [Cards Action (GitHub)](https://docs.mnemom.ai/concepts/cards-action.md): mnemom/cards-action@v1 — the GitHub Action that posts a care-framed PR comment whenever a .mnemom//**.yaml manifest changes. Editorial-feedback layer for the cards primitive. - [A2A AgentCard Export](https://docs.mnemom.ai/concepts/a2a-export.md): One-way projection of the canonical alignment card into the A2A AgentCard envelope, with the AAP attestation token attached as an A2A extension. - [AAP Attestation Tokens](https://docs.mnemom.ai/concepts/aap-attestation.md): Mnemom-signed JWS tokens that prove a canonical card's identity. Verifiable offline against /v1/.well-known/jwks.json, no Mnemom credentials needed. - [Transparency Log](https://docs.mnemom.ai/concepts/transparency-log.md): An append-only public log of every canonical card identity Mnemom has ever composed. Sigstore-Rekor-compatible row shape, Merkle inclusion proofs on demand. - [Alignment Cards](https://docs.mnemom.ai/concepts/alignment-cards.md): Declarative specifications of an AI agent's intended values, boundaries, and behavioral commitments - [Protection Card](https://docs.mnemom.ai/concepts/protection-card.md): The YAML document that tells Safe House how to defend an agent at runtime. Mode, thresholds, screen surfaces, trusted sources — all edited, composed, and audited like an alignment card. - [Card Composition](https://docs.mnemom.ai/concepts/card-composition.md): How alignment and protection cards compose across platform > org > (team)? > agent scope, with field-level rules, exemptions, and canonical-card semantics. ### Posture & Policy - [Trust Posture](https://docs.mnemom.ai/concepts/trust-posture.md): A named, versioned, cloneable, assignable oversight policy. Mnemom ships three opinionated defaults; orgs use them as-is or clone-and-customize. - [Trust Posture vs. Cards](https://docs.mnemom.ai/concepts/posture-vs-cards.md): Postures and cards are parallel artifact types: postures are team-scoped policy input; cards are agent-scoped runtime treatment. They cooperate through two well-defined join points. - [Mnemom-Shipped Default Postures](https://docs.mnemom.ai/concepts/mnemom-default-postures.md): Three opinionated defaults — Standard, High-Compliance, Low-Latency — Mnemom seeds at migration time. Use as-is or clone-and-customize. - [Posture Versioning](https://docs.mnemom.ai/concepts/posture-versioning.md): Postures are versioned with forward-only revision history. Every mutation creates a new revision; rollback creates a forward revision; old revisions stay immutable and queryable. - [Policy Engine](https://docs.mnemom.ai/concepts/policy-engine.md): Governance-as-code: capability mappings, forbidden rules, and enforcement modes, all declared in the alignment card's capabilities + enforcement sections - [CLPI: Card Lifecycle & Policy Intelligence](https://docs.mnemom.ai/concepts/clpi.md): The governance layer built on top of alignment cards: policy enforcement, violation reclassification and trust recovery, fleet intelligence, and checkpoint anchoring ### Detection & Safe House - [Safe House](https://docs.mnemom.ai/concepts/safe-house.md): Per-customer perimeter that screens every agent transaction at four checkpoints — front door, back door, inside.autonomy, inside.integrity — under the canonical off / observe / nudge / enforce enforcement modes. - [AP-Traces](https://docs.mnemom.ai/concepts/ap-traces.md): Structured records of agent actions — written at the integrity checkpoint (enforce gates same-turn; observe/nudge record post-hoc) and enriched afterward, then re-read longitudinally for drift and coherence analysis - [Integrity Checkpoints](https://docs.mnemom.ai/concepts/integrity-checkpoints.md): Per-turn analysis of LLM reasoning to detect prompt injection, manipulation, and drift - [Provider Support](https://docs.mnemom.ai/concepts/provider-support.md): Honest per-provider coverage of every Mnemom feature — what works, what's partial, and what's not supported across Anthropic, OpenAI, and Gemini, including Anthropic and OpenAI models hosted on Microsoft Foundry - [Drift Detection](https://docs.mnemom.ai/concepts/drift-detection.md): Detecting when AI agent behavior diverges from declared alignment over time - [Value Coherence](https://docs.mnemom.ai/concepts/value-coherence.md): Pairwise verification that interacting AI agents share compatible values - [Fleet Coherence](https://docs.mnemom.ai/concepts/fleet-coherence.md): Dimensional team coherence scoring — governance, conflict surface, structural invariants, and narrative helpers. No single fleet score. - [Sideband Detection](https://docs.mnemom.ai/concepts/sideband-detection.md): How Mnemom's observer detects coherence drops, fault lines, fleet partitions, and behavioral drift across teams of agents, and routes the findings to operators. - [Governance Signals](https://docs.mnemom.ai/concepts/governance-signals.md): Operator-actionable observations from Mnemom's platform detectors — fleet drift, value fault lines, coherence drops, behavior drift. Surfaced to operators (UI, webhook, REST). The platform never injects governance signals into an agent's prompt. - [Webhook Contract](https://docs.mnemom.ai/concepts/webhook-contract.md): Seven invariants every Mnemom webhook event carries — stable name, JSON Schema, HMAC signature, idempotency key, retry policy, replay path, surface separation. ### AEGIS - [AEGIS](https://docs.mnemom.ai/concepts/aegis.md): Adaptive Enforcement, Governance & Intelligence Substrate — the runtime screening layer of Safe House. Checks every gateway call at four checkpoints, under three per-agent settings, each set to one of four enforcement modes. - [Managed Rules](https://docs.mnemom.ai/concepts/managed-rules.md): Detection rules AEGIS delivers to every gateway as an Ed25519-signed rule set. Each gateway verifies the signature before it uses the rules and picks up changes within minutes. New rules are reviewed before they reach production. - [Substrate Fingerprint](https://docs.mnemom.ai/concepts/substrate-fingerprint.md): The provider, model, SDK and optional lockfile hash recorded on each integrity checkpoint, so you can tie an agent's decisions to the stack it ran on. ### Reputation & Risk - [Mnemom Trust Rating](https://docs.mnemom.ai/concepts/reputation-scores.md): The Mnemom Trust Rating™ — a credit score for AI agents. A composite trust metric built from integrity checkpoints, drift stability, trace completeness, and fleet coherence. - [Team Trust Rating](https://docs.mnemom.ai/concepts/team-reputation.md): The Team Trust Rating — a composite trust metric for teams of AI agents. Built from coherence history, member quality, operational record, structural stability, and assessment density. - [Risk Assessment](https://docs.mnemom.ai/concepts/risk-assessment.md): Context-aware, real-time risk scoring for individual agents and teams — with optional zero-knowledge proofs ### Lifecycle - [Card Lifecycle](https://docs.mnemom.ai/concepts/card-lifecycle.md): How alignment cards evolve as living artifacts — amendments, reclassification, trust recovery, and version tracking - [On-Chain Verification](https://docs.mnemom.ai/concepts/on-chain-verification.md): How Mnemom anchors reputation Merkle roots and score batches to Base L2, what that buys you, and its honest scope limits. ### Gateway - [Mnemom Gateway Overview](https://docs.mnemom.ai/gateway/overview.md): Transparent AI agent tracing gateway with AAP and AIP compliance - [CLI Reference](https://docs.mnemom.ai/gateway/cli.md): Mnemom CLI commands and usage - [mnemom agent](https://docs.mnemom.ai/gateway/agent.md): Launch a governed Claude Code session through the Mnemom gateway — install once, run one command, get an agent set up for you. - [Sign-in options](https://docs.mnemom.ai/gateway/agent-sign-in.md): How Claude Code signs in to Anthropic when you launch it with mnemom agent: a Claude Team or Enterprise subscription, or an Anthropic API key. - [Savings estimates](https://docs.mnemom.ai/gateway/agent-savings.md): What the mnemom agent first-run savings screen and the in-session savings counter show, how the estimate is worked out, and how to turn each one off. - [Context Fold for Every Agent](https://docs.mnemom.ai/gateway/context-fold.md): How the gateway folds long Anthropic conversations for every governed agent without an x-mnemom-conversation-id header, how the owner opts out on the alignment card, and what changes for the agent. - [Implicit goal mode](https://docs.mnemom.ai/gateway/implicit-goal-mode.md): Goal alignment with nothing written up front: the gateway builds the goal from your messages, updates it as you steer, and judges each step. - [Durable sessions (experimental)](https://docs.mnemom.ai/gateway/durable-sessions.md): Lock a named Mnemom agent session so a host daemon on your Mac keeps it running and brings it back after a crash or reboot, then steer it with send, questions, pause, resume, done, kill and amend. - [Slack channel](https://docs.mnemom.ai/gateway/slack-channel.md): Let agent sessions post to Slack and receive replies from the people you allow - [Enforcement Modes](https://docs.mnemom.ai/gateway/enforcement.md): Configure how the Mnemom Gateway responds to alignment and integrity violations - [Cross-Dialect Egress](https://docs.mnemom.ai/gateway/router-cross-dialect.md): Call the gateway with an Anthropic Messages request and route it to OpenAI — the gateway translates dialects in both directions, using your own OpenAI key for the upstream leg - [Microsoft Foundry](https://docs.mnemom.ai/gateway/microsoft-foundry.md): Route Claude and OpenAI models that you host on Microsoft Foundry through the Mnemom gateway, using your Foundry resource, your deployment and your credential, with Mnemom governance on every request - [Safe House Gateway Integration](https://docs.mnemom.ai/gateway/safe-house-overview.md): How Safe House integrates with the Mnemom gateway request pipeline — phases, caching, AIP enrichment, and attestation - [Custom Conscience Values](https://docs.mnemom.ai/gateway/conscience-values.md): Per-org alignment policies injected into the AIP conscience prompt - [Organization Card Templates](https://docs.mnemom.ai/gateway/org-card-templates.md): Define alignment and protection card policies that apply to all agents in your organization ### Pricing - [Pricing](https://docs.mnemom.ai/pricing/overview.md): Usage-based pricing in μ (1 μ = $0.01): gateway model calls at measured cost, a 5% inference share on requests from claimed and API-key agents with governance charged on top, fixed-price API products, and free non-model work ### Policy - [API Versioning Policy](https://docs.mnemom.ai/policy/versioning.md): Mnemom's canonical commitment for how the API evolves, how versions are negotiated, how long each version is supported, and how deprecations are signaled. ### Specifications - [Alignment Card Schema](https://docs.mnemom.ai/specifications/alignment-card-schema.md): Normative YAML schema for the unified alignment card, including every section, field type, required/optional status, and composition semantic. - [Attestation Token Schema](https://docs.mnemom.ai/specifications/attestation-token.md): Canonical wire format for AAP attestation tokens (v1) — JWS Compact serialization, EdDSA-signed payload. - [Transparency Log Schema](https://docs.mnemom.ai/specifications/transparency-log-schema.md): Canonical row shape for the append-only canonical-card transparency log (v1). Sigstore-Rekor-compatible mapping documented. - [Conduct Record](https://docs.mnemom.ai/specifications/conduct-record.md): The signed conduct record and PR record Mnemom builds for a recorded coding session: fields, Ed25519 signing, the digest chain, offline verification with @mnemom/verify, the two PR checks, and what a valid record does not prove. - [Protection Card Schema](https://docs.mnemom.ai/specifications/protection-card-schema.md): Normative YAML schema for the protection card, including every section, field type, required/optional status, and composition semantic. - [Team Template Schemas](https://docs.mnemom.ai/specifications/team-template-schemas.md): Normative JSON Schemas for team-scope alignment + protection templates. - [Trust Posture Schema](https://docs.mnemom.ai/specifications/trust-posture-schema.md): Normative JSONB body schema for Trust Posture revisions, plus the per-axis composition rules the composer uses to fold Platform → Org → Team. - [Pending Advisories Schema](https://docs.mnemom.ai/specifications/pending-advisories-schema.md): Wire format for pending_advisories rows + the closed source taxonomy and its compatibility rules. - [Governance Signals Schema](https://docs.mnemom.ai/specifications/governance-signals-schema.md): Table layout, RPC surface, and webhook event taxonomy for governance_signals. - [OpenTelemetry Attributes](https://docs.mnemom.ai/specifications/otel-attributes.md): Complete reference for AIP/AAP OTel span attributes and metrics - [Policy DSL Specification](https://docs.mnemom.ai/specifications/policy-dsl.md): Normative schema for the policy-engine's internal Policy shape — capability mappings, forbidden rules, escalation triggers, and defaults — as derived from an alignment card's capabilities + enforcement sections. - [agent-preview/v1 Specification](https://docs.mnemom.ai/specifications/agent-preview-v1.md): Normative spec for the agent-preview/v1 URL preview manifest: a structured JSON sibling for every URL, discoverable via . MIME application/vnd.agent-preview+json. Mnemom is the reference implementation; the spec is vendor-neutral. - [Managed Rule Envelope Schema](https://docs.mnemom.ai/specifications/managed-rule-envelope-schema.md): Wire format for the signed envelope the AEGIS promoter writes to two independent storage tiers and the gateway reads at request time. Documents the customer-readable surface; per-recipe internal columns are out of scope. ### Changelog - [Changelog](https://docs.mnemom.ai/changelog.md): Release history and updates for Mnemom protocols and infrastructure ## Guides ### Identity & Cards - [Agent Claim Flow](https://docs.mnemom.ai/guides/agent-claim-flow.md): How an owner adopts a gateway-provisioned agent into their org. Claim is the second half of the gateway-first lifecycle — it gives an unclaimed agent an owner and a real org. - [Authentication](https://docs.mnemom.ai/guides/authentication.md): Passkey sign-in, AAL2 step-up, session lifecycle, API key rotation, and account recovery for the Mnemom dashboard and API. - [Passkeys](https://docs.mnemom.ai/guides/passkeys.md): Sign in to Mnemom with a single biometric gesture — no password, no TOTP code. Passkeys use the WebAuthn standard and sync across your devices via iCloud Keychain, Google Password Manager, or 1Password. - [OAuth 2.0 Device Authorization Flow](https://docs.mnemom.ai/guides/oauth-device-flow.md): How headless agents authenticate with Mnemom using the OAuth 2.0 device-code grant (RFC 8628) — device-authorization request, user verification, token poll loop, and error handling. - [API Keys](https://docs.mnemom.ai/guides/api-keys.md): Personal vs. organization keys, capability scopes, and how to rotate them safely. - [Rotating Your Agent's API Key](https://docs.mnemom.ai/guides/agent-key-rotation.md): How to update an agent's provider key binding without losing history - [Multi-Agent Setup](https://docs.mnemom.ai/guides/multi-agent-setup.md): Configure multiple agents with Alignment Cards, run coherence checks, and monitor fleet alignment. - [Team Management](https://docs.mnemom.ai/guides/team-management.md): Create teams, manage rosters, derive alignment cards, and monitor team reputation - [Alignment Card Management](https://docs.mnemom.ai/guides/card-management.md): Create, validate, publish, and manage alignment cards for your AI agents - [Managing Card Composition](https://docs.mnemom.ai/guides/card-composition.md): Practical guide: manage org templates, grant exemptions, inspect canonical cards, and debug composition across the platform > org > agent hierarchy. - [Verify a Canonical Card Offline](https://docs.mnemom.ai/guides/verify-card.md): Use mnemom verify-card to prove a Mnemom-published canonical card was composed at the claimed identity, without holding any Mnemom credentials. - [Org Admin](https://docs.mnemom.ai/guides/org-admin.md): What an organization owner or admin can do, the roles in your org, and how to escalate the things this surface deliberately doesn't cover. - [Set up the Mnemom record](https://docs.mnemom.ai/guides/record-setup.md): Turn on the Mnemom record from the CLI, link one or more GitHub orgs, add more later, use it without GitHub, and see what it costs. - [Contracted agents](https://docs.mnemom.ai/guides/contracted-agents.md): Turn a GitHub repo on so that a labelled issue is sealed as a contract and an agent opens the PR, and choose where its jobs run: your CI or your own machine. - [Hand an issue to the contracted agent](https://docs.mnemom.ai/guides/contracted-agent-issues.md): Label a GitHub issue mnemom, write it so it seals on the first try, set limits in a mnemom block or over MCP, and steer the pull request the agent opens with reviews, comments and CI. - [Run Org Scan yourself](https://docs.mnemom.ai/guides/org-scan-run-it-yourself.md): Run an Org Scan from the command line or a GitHub Actions workflow, and read your Anthropic usage without a key through workload identity federation (WIF). - [Read the Org Scan value report](https://docs.mnemom.ai/guides/org-scan-value-report.md): What the Org Scan value report shows, how to read a row, what is measured and what is an estimate, what Mnemom keeps, and how to export the report, open it again and verify its signature. ### Posture & Policy - [Posture Cloning Workflow](https://docs.mnemom.ai/guides/posture-cloning.md): Clone a Mnemom-shipped default and tighten it for your fleet. The clone-and-customize pattern preserves the upstream baseline, gives you your own forward-only revision history, and produces a clean compliance audit trail. - [Policy Management](https://docs.mnemom.ai/guides/policy-management.md): Create, test, publish, and manage governance policies for your AI agents - [CI/CD Policy Gates](https://docs.mnemom.ai/guides/ci-cd-policy-gates.md): Integrate policy validation and evaluation into your CI/CD pipeline with GitHub Actions and GitLab CI - [Trust Recovery](https://docs.mnemom.ai/guides/trust-recovery.md): Recover agent trust scores by reclassifying card-gap violations and amending alignment cards - [Explain and Remediate](https://docs.mnemom.ai/guides/explain-and-remediate.md): Use the explain endpoint to understand why an agent's policy evaluation produced a particular verdict, then act on the structured remediations to resolve each finding. - [Simulate Before Commit](https://docs.mnemom.ai/guides/simulate-before-commit.md): Use the simulate endpoint to dry-run gateway + observer evaluators against a hypothetical input or tool call. Verify a tool call would clear policy before you let an agent invoke it in production. - [Compliance Attestation Foundation](https://docs.mnemom.ai/guides/compliance-attestation-foundation.md): Use versioned Trust Postures as audit evidence for SOC 2, EU AI Act, HIPAA, and ISO 27001. The revision history API answers 'what was our detection policy on date X'. ### Detection & Safe House - [Observability with OpenTelemetry](https://docs.mnemom.ai/guides/observability.md): OpenTelemetry exporter for AIP integrity checkpoints and AAP verification results - [Card-Change Notifications](https://docs.mnemom.ai/guides/card-change-notifications.md): Subscribe to SSE or webhook notifications when a canonical card changes, sourced from the transparency log. A separate, per-agent mechanism from org-level webhooks. - [Tuning Sideband Detection via Trust Posture](https://docs.mnemom.ai/guides/sideband-detection.md): Tune coherence, fault-line, and fleet detectors via Trust Posture. Enable an axis, set a cadence, pick a severity, and the observer applies the strictest-wins fold of Platform → Org → Team. - [Operating Governance Signals](https://docs.mnemom.ai/guides/operating-governance-signals.md): Operator runbook for ack/resolve/dismiss workflow, notification destinations (Slack/email/PagerDuty/webhook), and escalation rules. - [Webhook Notifications](https://docs.mnemom.ai/guides/webhooks.md): Create org-level webhook endpoints, verify signed deliveries, and handle retries — real-time HTTP POST notifications instead of polling. - [Security & Trust Model](https://docs.mnemom.ai/guides/security-trust-model.md): How Mnemom handles data, authentication, encryption, and threat modeling across the alignment infrastructure. - [Supply Chain Trust](https://docs.mnemom.ai/guides/supply-chain-trust.md): How Mnemom publishes its npm packages, which carry SLSA provenance, and how to verify the packages you install. - [Opt in to lockfile-hash substrate attribution](https://docs.mnemom.ai/guides/lockfile-hash-opt-in.md): Set the X-Mnemom-Lockfile-Hash header on outbound requests so each integrity checkpoint records which deployed dependency graph produced it. SHA-256 digest of your resolved manifest; the hash is sent, the manifest is not. - [Protection Card Management](https://docs.mnemom.ai/guides/safe-house-config.md): Author, publish, validate, and tune the protection card that configures Safe House for an agent. - [Safe House Threat Model](https://docs.mnemom.ai/guides/safe-house-threat-model.md): What Safe House protects against — specific attack vectors, detection mechanisms, known limits, and false positive management. - [OWASP Agentic Top 10 Mapping](https://docs.mnemom.ai/guides/owasp-agentic-top-10.md): How Safe House's shipped threat patterns map to the OWASP Top 10 for Agentic Applications (December 2025). Coverage mechanisms, gaps, and what to pair with application-layer controls. - [Safe House Monitoring](https://docs.mnemom.ai/guides/safe-house-monitoring.md): Monitor Safe House in production — the Security Observatory, key metrics, session escalation, campaign detection, pattern promotion, and evaluation history queries. ### Reputation & Risk - [Improving Your Mnemom Trust Rating](https://docs.mnemom.ai/guides/improving-reputation.md): A practical guide to building, monitoring, and recovering your Mnemom Trust Rating™ — component-by-component strategies for agent builders and operators. - [Risk Engine](https://docs.mnemom.ai/guides/risk-engine.md): How to assess, gate, and monitor risk for individual agents and teams using the Mnemom Risk Engine - [Fault Line Analysis](https://docs.mnemom.ai/guides/fault-line-analysis.md): Identify, classify, and resolve value misalignments in multi-agent teams. Fault-line classification runs alongside v2 coherence scoring and produces an actionable triage list. - [Embeddable Trust Badges](https://docs.mnemom.ai/guides/reputation-badges.md): Add trust signals to your README, website, Agent Card, or documentation with dynamic SVG badges showing your agent's Mnemom Trust Rating™. ### Lifecycle - [API Versioning](https://docs.mnemom.ai/guides/api-versioning.md): How to pin your integration to a specific API version and manage upgrades safely. - [Compliance Posture](https://docs.mnemom.ai/guides/compliance.md): Honest status across the regulatory frameworks enterprise buyers ask about — EU AI Act, SOC 2, HIPAA, FedRAMP, and more. - [EU AI Act Compliance](https://docs.mnemom.ai/guides/eu-compliance.md): EU AI Act Article 50 transparency mapping (AAP + AIP) and the NIST AI RMF (GOVERN/MAP/MEASURE/MANAGE) mapping to Mnemom's shipped controls - [GDPR Right to Erasure](https://docs.mnemom.ai/guides/gdpr-data-subject-rights.md): Right to erasure (Article 17) — how Mnemom handles agent deletion requests - [Aletheia Customer Voice — Privacy & Data Use](https://docs.mnemom.ai/guides/customer-voice-privacy.md): What Aletheia Customer Voice collects, how it is used, retention periods, and your rights under GDPR - [Aletheia Customer Voice — Retention Policy](https://docs.mnemom.ai/guides/customer-voice-retention.md): Retention settings, on-request deletion, and audit trail for Aletheia Customer Voice data - [WEF Agent Governance Framework](https://docs.mnemom.ai/guides/wef-governance.md): Mapping AAP and AIP to the World Economic Forum's AI agent governance framework - [NIST CSF / 800-53 Control Mapping](https://docs.mnemom.ai/guides/nist-csf-800-53-mapping.md): How Mnemom's shipped controls map to the NIST Cybersecurity Framework 2.0 functions and NIST SP 800-53 Rev. 5 control families. Coverage mechanisms named explicitly; gaps stated as gaps. - [SLA and Incident Response](https://docs.mnemom.ai/guides/sla-and-incident-response.md): The public status page, incident severity classification, and how Mnemom communicates during an incident. Contractual uptime/RTO/RPO numbers live in your Enterprise MSA. - [Upgrading AAP & AIP](https://docs.mnemom.ai/guides/upgrading-to-1-0.md): Migrating to the current AAP (2.0.0) and AIP (1.3.0) SDK releases, including the 1.0.0 stability commitment and the 2.0.0 card-shape breaking change. - [Upgrading to AAP 0.5.0](https://docs.mnemom.ai/guides/upgrading-to-0-5.md): Step-by-step migration guide for updating your agents, cards, and policies to AAP 0.5.0 - [Using On-Chain Verification](https://docs.mnemom.ai/guides/on-chain-verification.md): The three read-only endpoints for checking an agent's on-chain anchor status, verifying its Merkle proof, and browsing anchor history. ## Protocols ### Protocols - [Protocol Overview](https://docs.mnemom.ai/protocols/overview.md): Understanding AAP, AIP, and CLPI — the protocol stack for agent trust ### Agent Alignment Protocol - [AAP Specification](https://docs.mnemom.ai/protocols/aap/specification.md): Full protocol specification for the Agent Alignment Protocol (AAP) -- the unified card shape shipped in SDK v2.0.0 - [AAP Quick Start Guide](https://docs.mnemom.ai/protocols/aap/quickstart.md): Get your agent producing verifiable alignment traces in 5 minutes - [AAP Architecture](https://docs.mnemom.ai/protocols/aap/architecture.md): System architecture of the Agent Alignment Protocol including component relationships, data flow, and extension points - [AAP Security Model](https://docs.mnemom.ai/protocols/aap/security.md): Security model for the Agent Alignment Protocol including threat analysis, cryptographic requirements, and defense in depth - [AAP Limitations and Honest Claims](https://docs.mnemom.ai/protocols/aap/limitations.md): What the Agent Alignment Protocol does and does not provide -- normative limitations that implementers must communicate - [AAP Calibration Methodology](https://docs.mnemom.ai/protocols/aap/calibration.md): How AAP's drift detection thresholds were empirically derived from multi-turn agent conversation analysis - [A2A Integration Guide](https://docs.mnemom.ai/protocols/aap/a2a-integration.md): How to extend A2A Agent Cards with AAP alignment properties for value coherence checks before agent-to-agent coordination - [Adding AAP to MCP Tools](https://docs.mnemom.ai/protocols/aap/mcp-migration.md): How to extend MCP tools and servers with AAP alignment properties for alignment verification of tool invocations ### Agent Integrity Protocol - [Agent Integrity Protocol (AIP) Specification](https://docs.mnemom.ai/protocols/aip/specification.md): Complete protocol specification for per-turn integrity assurance of autonomous agents via thinking block analysis - [Integrity Certificates](https://docs.mnemom.ai/protocols/aip/certificates.md): Machine-readable cryptographic certificates that bundle all verification evidence for an integrity checkpoint into a single, self-describing document. - [Verifiable Verdicts](https://docs.mnemom.ai/protocols/aip/verifiable-verdicts.md): Zero-knowledge proofs for verdict derivation, enabling cryptographic verification that an integrity verdict was computed correctly from its inputs. - [AIP Quickstart](https://docs.mnemom.ai/protocols/aip/quickstart.md): Get from zero to per-turn integrity checking in 5 minutes - [AIP Security Model](https://docs.mnemom.ai/protocols/aip/security.md): Threat model and security considerations for the Agent Integrity Protocol - [AIP Limitations](https://docs.mnemom.ai/protocols/aip/limitations.md): What AIP does and does not do -- read before depending on AIP for anything consequential ## OpenAPI Specs - [openapi](/api-reference/openapi.json) - [minimal-openapi](/scripts/test-fixtures/minimal-openapi.json) ## Optional - [GitHub](https://github.com/mnemom) - [Blog](https://www.mnemom.ai/blog/) > The links below point to documentation indexes. Follow each `/_llms/` index recursively until you reach documentation pages. ## Indexes - [English / API Reference (686 pages)](https://docs.mnemom.ai/_llms/en/api-reference.txt): API Reference, Auth, OAuth, Agents, Alignment, Protection, Postures, Unified cards, Enforcement, Governance, Risk, Drift, Integrity, Checkpoints, Verification, Reputation, Team Reputation, Teams, Organizations, Licensing, Billing, Tools, Catalog, Notifications, Webhook Notifications, AIP Webhooks, Transparency, Attestation, A2A, Recipes, Sideband, Traces, Analyze, Reclassification, Intelligence, Policy, On-Chain, Transactions, Conscience Values, Agent Containment, Agent Runtime, Safe House, Trust & Network (AEGIS), Agent Groups, Aletheia Support, Mu, Code, Coherence, Consent, Dojo, Domains, Feedback, Webhooks, Invite Message Template, Invites, MCP Servers, Onboarding, Presentations, Blog. - [English / API Reference / Alignment (82 pages)](https://docs.mnemom.ai/_llms/en/api-reference/alignment.txt): Agent scope, Org scope, Team scope, Platform scope, General. - [English / API Reference / Protection (50 pages)](https://docs.mnemom.ai/_llms/en/api-reference/protection.txt): Agent scope, Org scope, Team scope, Platform scope, General. - [English / API Reference / Organizations (76 pages)](https://docs.mnemom.ai/_llms/en/api-reference/organizations.txt): Create organization, List organizations for authenticated user, Get organization details, Update organization, Delete organization (owner only), List organization members, Update member role (owner only), Remove member or self-leave, Invite member to organization, List pending invitations, Revoke a pending invitation, Accept an organization invitation, List agents in organization, Create organization API key, List organization API keys, Revoke organization API key, Rotate an organization API key, Get SSO configuration for organization, Configure SAML SSO for organization (owner only), Remove SSO configuration (owner only), Test SSO metadata URL validity, Enable or change the Safe House posture for an org, DEPRECATED — use GET /v1/alignment/org/{org_id}., DEPRECATED — use PUT /v1/alignment/org/{org_id}., DEPRECATED — use DELETE /v1/alignment/org/{org_id}., DEPRECATED — use GET /v1/protection/org/{org_id}., DEPRECATED — use PUT /v1/protection/org/{org_id}., DEPRECATED — use DELETE /v1/protection/org/{org_id}., Daily request + token totals for the org, Per-model cost rollup for the org, Read-only billing summary for the org, Open the Stripe Customer Portal for this org, Read the org audit log, Upload organization avatar, N-Way value coherence for the org fleet (with care-framed recommendations), Export a snapshot of the org's agent fleet, GitHub App Setup URL — redirect back to the originating console cell, Erase all this org's Customer Voice feedback (GDPR), Export this org's Customer Voice feedback (GDPR), Read this org's Customer Voice retention settings, Update this org's Customer Voice retention settings, Erase one Customer Voice feedback submission (GDPR), Read this org's feedback, GitHub installation status for organization (member), Disconnect GitHub installation (owner/admin), Bind a GitHub installation to the organization (owner/admin), Begin GitHub App installation (owner/admin), Create a 24-hour repository token grant (member), Revoke a token grant (member or cell service), Mint a 1-hour installation token under a grant (member or cell service), List repositories visible to the org's GitHub installation (member), Self-serve restore a tombstoned organization (owner or admin, within 30 days of tombstoning), Per-person, per-model consumption for the org, Who holds anything in the organization, from grant. (owner/admin/auditor), What one subject holds in the organization, from grant. (owner/admin/auditor), What the caller holds in the organization, from grant. (any member), Grant settings for organization (owner/admin/auditor), Org scan GitHub App Setup/Callback URL — redirect back to the originating page, Theater board live feed, replay window and polling fallback (TH-G-06), List the leases the signed-in person may see (org admin, or the agent's sponsor), Stop every live lease of one agent, with a reason (org admin, or the agent's sponsor), One lease receipt: agent, sponsor, scope, issued/expires, state, Stop one lease, with a reason (org admin, or the agent's sponsor), The open widening requests the signed-in person may decide (any org member), Allow or Refuse a widening request, as the signed-in person (any org member), Org scan GitHub App connection status (member), Disconnect the Org scan GitHub App (owner/admin), Bind an Org scan GitHub App installation to the organization (owner/admin), Begin the Org scan GitHub App installation (owner/admin), List repositories the org's Org scan GitHub App can see (member), Read a repo's record App settings, Change a repo's record App settings, Report a session commit to the record commit index, Read a conversation's signed conduct record, Read a conversation's model cost and µ totals, Issue a signed share card over the org's own signed statements. - [English / API Reference / Dojo (44 pages)](https://docs.mnemom.ai/_llms/en/api-reference/dojo.txt): Mint a single-use claim-intent ticket (anonymous, birth-token proof)., Exchange a claim-intent ticket — claim + bind (authed)., Mint an ephemeral compute/operate key (gateway + api:read + api:write)., List the caller's dojo agents + dojo_enabled flag., List configured dojo provider keys (admin-only, non-secret metadata)., Set the dojo provider API key for a provider (admin-only)., Verify a Dojo completion receipt against the published JWKS., Issue the signed completion receipt for a finished Dojo run., Per-run Safe House detector posture (staff-only)., List Dojo scenarios for the How-it-works modal., Get one Dojo scenario by slug (How-it-works modal)., List Try-Me Configs., Mint a Try-Me Config (scenario + model) with a server-generated token., Batch-mint Try-Me Configs and email each invitee an invitation., Burn/expire a Try-Me Config, freeing its token., Re-send an invitation email for an already-minted Try-Me Config., Re-arm a Try-Me Config, making its token resolvable again., Record the born agent on its Try-Me Config (authed, owner-only)., Self-serve re-arm of a stuck (burned) demo token., Self-serve re-arm of a stuck demo token (POST form)., Resolve a single-use Try-Me token to its onboarding briefing manifest., Resolve a single-use Try-Me token to its briefing manifest (POST form)., Force-end every active sim session for an agent (runaway recovery)., Create a Dojo sim session and begin the runner ("Begin Sim")., Submit a Manual-mode Agent Smith attack., Toggle raw-payload capture for a Dojo run (Mnemom staff)., Update a running Dojo sim session's live config (mode / security posture / judge)., Export a full Dojo run to a downloadable file., Keep an Auto Dojo sim session alive (liveness lease)., Abort a running Dojo sim session., Server-Sent Events stream of the Dojo sparring narrative., Get the per-scenario Agent Smith sparring config (admin console)., Upsert the per-scenario Agent Smith sparring config (admin console)., Agent Smith coverage grid (admin console)., Agent Smith arms-race co-evolution timeline (briefing)., Agent Smith live evolution generation (admin console)., Agent Smith matchup profile (Evo badge)., Agent Smith promotion-run audit feed (admin console)., Agent Smith per-scenario resistance trend (admin console)., List Agent Smith review-queue candidates with provenance (admin console)., Promote (or merge) an Agent Smith review candidate (admin console)., Reject an Agent Smith review candidate with a reason (admin console)., List Agent Smith catalog techniques (admin console)., Get one Agent Smith technique by id or slug (admin console).. - [French (6 pages)](https://docs.mnemom.ai/_llms/fr.txt): Démarrage rapide. - [Spanish (6 pages)](https://docs.mnemom.ai/_llms/es.txt): Inicio rápido. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.