Scryo / Legal
SCRYO PRIVACY POLICY
How information is handled in Scryo.
1. Introduction and scope
Loci Labs AI, LLC (“Loci Labs,” “we,” “our,” or “us”) operates Scryo, a security research and intelligence platform for discovering, observing, and comparing publicly accessible Model Context Protocol (“MCP”) endpoints and related services. This Policy explains what personal information we handle, its sources and purposes, disclosures, retention, and privacy rights.
For the independent purposes described in this Policy, Loci Labs determines why and how information is processed and acts as a data controller where that term applies. This Policy covers users and people whose personal information appears in observations even if they have never created a Scryo account. Public availability does not by itself remove privacy protections.
This Policy applies to public pages, access applications and invitations, the authenticated application, API access where enabled, observations, communications, support, and billing where enabled. It should be read with the Terms of Service at https://scryo.loci-labs.ai/terms. Independent identity, payment, and endpoint providers may also have their own notices.
Scryo's shared technical intelligence is separate from private account, search, admission, and request records. An eligible endpoint submitted for observation and its resulting technical observations may become visible to other authorized users; the submitter's identity and private request-status records do not become part of that shared observation.
2. Categories of information
2.1 Access applications and invitations
When you request access, we process the name, email, organization, and intended use you provide, submission time and identifiers, application status, review and admission decisions, proposed role, invitation records, and notices sent. Invitations may also be issued without an access application. We process relevant issuance, expiry, revocation, and redemption information and limited anti-abuse and rate-limit information. The record-retention periods in Section 8 do not determine how long an invitation remains valid.
2.2 Accounts and sign-in
We process information needed to identify and authorize an account, including a verified email, account or identity-provider identifier, and an optional profile name or image where supplied. Clerk manages sign-in, authentication challenges, sessions, and profile/security settings. Depending on enabled options and the method chosen, it may interact with Google, Apple, GitHub, or an email verification service. Scryo separately maintains account identifiers, roles, access/admission status, permissions, Terms-acceptance records, and security audit records.
At signup, the intended combined declaration accepts the Terms and confirms the user is at least 18. We do not collect date of birth or government-issued ID for that self-declaration. We retain limited evidence of the accepted Terms version and acceptance time, subject to Section 8. Scryo support does not request passwords or multi-factor recovery codes.
2.3 Searches, observation requests, and private submission groups
We process queries and application requests to provide the requested features. We do not maintain a separate account-linked search history. Query content may be retained only where necessary in routine troubleshooting logs, under the 30-day rule in Section 8.
Private observation-request records may include the submitting account, endpoint URL or hash, local and upstream request identifiers, submission/status times, quota information, and outcomes. They support ownership, request status, limits, and security. Searches do not themselves trigger new observations; private request pages and requester identity are not shared intelligence.
The intended private submission-group design uses a separate randomly generated identifier to connect observations submitted by the same person. It does not reuse an account UUID, Clerk identifier, or an identifier derived from email. The account-to-group mapping is kept separately with restricted access and deleted with the account. The remaining private grouping supports relationships among submissions, not deliberate recovery of the former account's identity. We do not claim that deleting the mapping automatically makes every remaining record anonymous; applicable privacy rights and the retention assessment still apply.
Scryo and its infrastructure providers may also process IP addresses, browser/device information, request times, session/security identifiers, errors, and security events to operate and protect the service. Temporary rate-limit counters and logs have separate retention rules.
2.4 Feedback, support, and communications
For feedback or support, we process the submitted message, category, optional reply address, relevant limited application/browser context, correspondence, and case status. Email infrastructure may process email addresses, headers, bodies, attachments, and delivery/security metadata. We use this information to respond, troubleshoot, investigate abuse, and improve the service. Do not send passwords, access tokens, private investigation results, or unnecessary sensitive information; do not send government-ID copies for privacy verification.
Service emails include application acknowledgments, invitations and decisions, authentication messages, necessary operational notices, and material privacy-policy change notices. We do not currently run marketing campaigns or use email-open or click tracking.
2.5 Payments and paid API access
Beta features designated as free have no usage fee. API access is intended to be paid when launched; no downloadable dataset-export feature is planned. This description does not mean live payments or paid API access have been activated.
When paid billing is enabled, Stripe will process checkout and payment information through its hosted services. Scryo may receive or store a customer identifier and account binding, plan/status and billing-period information, checkout/invoice references, limited transaction outcomes, and event identifiers needed for entitlements and reconciliation. We do not intend to store full payment-card numbers in Scryo's application database. Payment information entered into Stripe is handled by Stripe under its notice.
2.6 Publicly discovered service information
The shared dataset may include endpoint URLs/domains; advertised service names and vendor/publisher claims; discovery sources and references; observation times; IP/DNS information; TLS certificate details; limited HTTP, authentication, and payment metadata; advertised MCP tools, resources, resource templates, and prompts and their names/descriptions; hashes/fingerprints; and changes or relationships among endpoints. Sources include public registries, directories, repositories, websites/documentation, and bounded requests to public Internet services.
Some metadata, such as a name, personal email, certificate subject, personal domain, or advertised contact information, may identify an individual. Scryo does not seek private credentials, protected resource contents, private content, or personal information unrelated to describing a public service. We review substantiated personal-information concerns and take targeted deletion, redaction, correction, or restriction action as described below. We do not represent that technical metadata cannot contain personal information.
3. Purposes and legal grounds
We use information to review applications; administer accounts and invitations; authenticate users and enforce permissions; provide searches, observations, technical history, relationships, and authorized API access; manage communications, support, and enabled billing; diagnose errors and availability; investigate security incidents, impersonation, fraud, and misuse; protect public endpoints from abusive requests; maintain accurate intelligence; meet legal obligations; and establish, exercise, or defend legal rights.
Where EU/UK GDPR or similar law applies, the ground depends on the activity. Contract performance or requested pre-contract steps supports eligible account/application and requested-service processing. Where the customer is an organization, legitimate interests may support administering its users' authorized access.
Security, reliability, support, and public-service intelligence rely, where applicable, on legitimate interests in protecting systems and users, investigating threats/fraud, and helping users assess public AI-facing services and their technical history. Processing personal information for these purposes must be necessary and proportionate, taking account of affected individuals' rights. Public availability and a user's acceptance of the Terms are not substitutes for that assessment.
We rely on legal obligations where processing is required by law and obtain consent where required for a particular activity. We do not use social sign-in information for advertising or include private account records in endpoint intelligence. Materially different purposes require updated disclosures and any consent required by law.
4. Shared observations and disclosure boundaries
Accepted endpoint observations, history, provenance, and technical relationships may be made available to authorized users and, when launched, paid API customers. Permitted downstream use is governed by the Terms: internal security work and limited excerpts in client reports, published research, and conference presentations. Resale or incorporation into customer-facing products requires a separate agreement. The Terms, not this Policy, define the license and attribution requirements.
API clients can save responses even though Scryo does not offer a downloadable dataset-export feature. Permitted reports or publications can reach people outside Scryo. A later correction, restriction, or deletion does not necessarily recall independently held copies. We will take required recipient-notification and other legal steps, and the Terms impose the agreed privacy-removal obligations on API customers for copies they control. We do not guarantee complete recall of independent third-party copies.
The shared record is intended to describe the endpoint, not the requester. We do not include private requester identities, search terms, admission decisions, feedback, request-status pages, or private submission-group identifiers in the shared corpus. Do not submit secret-bearing, private, personal, or otherwise unauthorized endpoints; the accepted URL and advertised metadata may be shared.
Historical observations are a core feature. Account closure or an endpoint operator's request does not automatically erase an endpoint or unrelated technical history. Personal-information rights and justified corrections, redactions, restrictions, and removals remain available. Technical relationships are not proof of identity, ownership, intent, or wrongdoing.
5. Recipients and other disclosures
We disclose information only as reasonably needed for the purposes described here to hosting, storage, networking, security, and communications providers, including AWS and Cloudflare; Clerk and selected sign-in providers; Stripe when paid billing is enabled; authorized personnel and contractors with a need for access and appropriate obligations; authorized users/API customers for designated shared observations only; and a successor or transaction counterparty in a merger, restructuring, financing, or asset transaction, subject to applicable law.
We may also disclose information as reasonably necessary to comply with law or legal process, respond to lawful government requests, protect rights/safety/security, enforce the Terms, or investigate incidents, abuse, or fraud. We may publish aggregated or properly de-identified statistics that do not reasonably identify an individual.
We do not provide private account, sign-in, or support information to advertising networks or data brokers. Paid access to shared technical observations is a separate activity and may incidentally involve personal information. We will provide the disclosures and choices applicable to that activity; this Policy does not make an unconditional representation that every intelligence disclosure falls outside laws governing sales or sharing of personal information.
6. Sign-in provider information
For Google sign-in, the intended OpenID Connect permissions are limited to openid, email, and profile, used through Clerk for identity verification, access, and optional basic profile display. Scryo's sign-in flow does not request Gmail messages, Drive files, contacts, or calendars.
Apple, GitHub, or email-code sign-in shares information according to the chosen provider and permissions. We use the verified identity for Scryo access, not to infer ownership of observed endpoints. Provider permissions can be managed with the provider; Scryo-related requests can be sent to support.
7. Cookies and similar technologies
Scryo and its identity/security providers use necessary cookies, tokens, or similar technologies for authentication, session management, short-lived invitation continuation, security, and basic functionality. Cloudflare and Clerk may set authentication or security cookies according to their configurations. Blocking necessary cookies may prevent sign-in or other essential features.
We use operational monitoring and error logs, not third-party behavioral analytics, advertising pixels, heatmaps, or session replay. Any future nonessential technology requiring consent will require an appropriate choice before use. Acceptance of the Terms is not substituted for a legally required separate consent.
8. Retention and deletion
The following schedule describes the intended retention design. We retain personal information only as reasonably necessary for its stated purpose, subject to applicable legal rights and narrowly justified retention exceptions.
Accounts and approved applications: retain account information and approved application/approval records while the account exists, then delete them on a valid account-deletion request. Redeemed invitation records follow the same account-linked rule. The internal target for routine active-system deletion is 28 calendar days from receipt of the request, subject to applicable legal deadlines, necessary verification, and valid retention exceptions. Confirmation does not silently restart that target. Limited acceptance, billing, and legally necessary compliance evidence are separate categories, not a reason to retain the full account.
Authentication and authorization audits: one year from each event, or deletion with the associated account if earlier, under the same active-system deletion target. This rule is distinct from routine troubleshooting logs and backup history.
Private operational observation-request records: approximately 30 days in active request stores, with cleanup during subsequent operations. The separate private submission-group history and account mapping follow the rule below; operational request expiry does not establish deletion of all backup or log copies.
Private submission-group history: no routine age-based expiry for the private relationship among submissions. Delete the account-to-group mapping with the account. Keep the grouping private and restrict access. Assess any personal information remaining in the grouping separately; indefinite technical-history retention does not override correction, objection, restriction, or deletion rights or justify keeping unnecessary identifying information. Do not retain other records to reconstruct the deleted mapping.
Shared technical intelligence: retain endpoint observations, provenance, changes, and history indefinitely, subject to targeted correction, redaction, restriction, de-identification, or deletion where appropriate or legally required. Continued retention of particular personal identifiers and continued disclosure to users must each remain justified; useful unrelated technical history may remain where lawful.
Feedback and ordinary support: retain in-product bug reports, feature suggestions, support cases, and related messages while open, then delete active records 90 days after case closure. The clock is closure, not initial submission or conversation creation. Privacy requests and required compliance evidence must not be incorrectly deleted merely because an ordinary support period expires.
Rejected applications: delete active application and rejection-decision records 90 days after rejection. Pending applications still awaiting a decision: close and delete active records 90 calendar days after submission.
Unused invitations: delete expired invitation records 90 days after expiration and revoked invitation records 90 days after cancellation. These periods do not extend invitation validity. Invitation records used to create an account are kept with that account and deleted with it.
Routine application/error logs: 30 calendar days from each entry. Query content is included only when necessary for troubleshooting; there is no separate account-linked search-history store. Temporary rate-limit counters are deleted once the enforcement window ends and they are no longer needed to enforce that limit. Other provider or system records require classification under the relevant purpose and schedule, not a default of indefinite retention.
Terms-acceptance evidence: retain only the accepted Terms version, acceptance time, and necessary evidentiary identity information until five years after account closure or the end of the relevant contract, whichever is later. Keep this limited evidence separate from the full account and private submission-group mapping. Delete it when the period ends unless a specific, documented legal preservation requirement applies; do not use it to reconstruct a deleted account-to-submission-group link.
Billing evidence when paid service launches: retain necessary invoice and transaction information for accounting, tax, and disputes for seven years after the relevant financial year ends. A longer period requires a documented applicable requirement. Keep these limited records separate from the full account and private submission groups, and do not use them to reconstruct a deleted account-to-submission-group link. Full payment-card data is not intended to be stored in Scryo's application database.
Privacy-request compliance evidence: retain a separate minimal compliance record while a request is open and for 24 months after request closure, subject to a different period required by applicable law or a documented legal preservation requirement. Record only what is necessary to demonstrate receipt, the nature and method of the request, verification outcome, response and action dates, outcome, and any permitted refusal or retention. Keep this record separate from ordinary support-message bodies, whose usual retention is 90 days after case closure. Do not retain the deleted dataset or a recoverable account-to-submission-group mapping as compliance evidence.
Recovery backups: the intended maximum history for backups containing account information is 90 days measured from each backup's creation, including versions and recovery copies. Expiring a snapshot does not delete the live database; data still live may appear in newer backups. After active-system deletion, residual backup information is kept out of normal use and expires under the backup schedule. Approved deletions must be reapplied before restored data is returned to service. This is a backup-age limit, not a promise that all copies disappear 90 days after a request. A specific legal preservation requirement must be assessed and narrowly documented separately.
Exceptions and holds: retain only information necessary for a specific applicable legal, accounting, security, anti-abuse, or evidentiary requirement, restrict access/use, document the basis and release condition, and remove it when that need ends. These exceptions do not authorize routine preservation of the full account, reconstruction of a deleted account-to-submission-group mapping, or automatic extension of every schedule. The expressly scheduled minimal acceptance, billing, and privacy-request evidence may retain only the limited identity information necessary for its stated purpose.
9. Security
We use organizational and technical measures designed to protect information proportionate to risk, including access controls, identity checks, multi-factor authentication where configured, encryption in transit and at rest where configured, role-based authorization, restricted administration, logging, and recovery backups. No service guarantees perfect security. Contact support@loci-labs.ai promptly about exposure or account compromise.
10. International handling
Scryo is operated by a U.S. company. Infrastructure and providers may process information in the United States and other countries with different privacy protections. The intended audience includes international users, including the EU and UK where legally permitted.
For a disclosure subject to international-transfer requirements, safeguards must match the actual recipient, role, destination, and processing arrangement. Contact support@loci-labs.ai for information about applicable safeguards. No vendor certification is assumed to cover every Loci data flow.
11. Privacy rights and requests
Depending on residence and applicable law, rights may include access/knowledge, a copy of information, correction, deletion, objection or restriction, withdrawal of consent where used, portability, and opt-out of qualifying sales, targeted advertising, or other covered processing. We will not unlawfully discriminate for exercising applicable rights.
Loci Labs AI, LLC is responsible for receiving, tracking, and responding to requests at support@loci-labs.ai. Describe the right, relevant account or observation, and action requested. Verification must be proportionate to the request and risk. We do not request government-issued identification, including as a fallback or through a provider acting for us; do not send such documents.
For deletion of your own Scryo account, the standard confirmation is a one-time verification link emailed to the address already registered on the account. If you no longer control that email, company support will review other evidence manually, without government ID. This does not authorize an unverified person to delete another account. Verification for other rights, nonusers, or authorized representatives is assessed according to the request and available evidence, without requiring a nonuser to create an account.
If personal information about you appears in an observation, provide its URL or enough information to locate the record and explain your connection to the information. Company support will make every reasonable effort to verify the connection and delete the specified personal information, including redacting it from retained technical records. Where law requires action, the duty is not limited to best efforts. Accuracy of an underlying technical fact alone does not determine whether an associated personal identifier may be retained or disclosed.
For personal information approved for removal, the intended process filters that specific information from future observations and reprocessing rather than automatically stopping observation of an entire endpoint. Removals must be preserved through restoration. We will explain the action's scope and do not guarantee prevention of every possible recurrence or recall of every independent copy. Broader deletion, restriction, or cessation required by law takes priority.
We will respond within applicable legal timeframes, explain any permitted denial or limitation, and provide applicable appeal or complaint information. Where available, appeal by replying to our determination with “Privacy appeal.” EEA/UK individuals may complain to the competent data-protection authority. Account closure does not itself mean immediate erasure of all independently collected technical history or every protected backup; the distinctions in Section 8 apply without reducing statutory rights.
12. Children
Scryo accounts are for people at least 18 years old and the service is not directed to children. The combined signup declaration in the Terms confirms eligibility without collecting date of birth or government-issued ID for that purpose; it is self-declaration, not independent age verification. If you believe a minor has submitted personal information or created an account, contact support so the company can review and take appropriate action.
13. Changes to this Policy
We will post revised policies at https://scryo.loci-labs.ai/privacy and update the effective/last-updated date. For material changes to how personal information is collected, used, or shared, we will also email registered users. Any further notice or consent required by law must be provided. Earlier draft acceptance does not authorize materially different future practices where a new choice is required.
14. Contact
Loci Labs AI, LLC
Privacy and legal inquiries: support@loci-labs.ai
Mailing address: 8401 Mayland Dr Ste A, Richmond, VA 23294-4648, United States
Website: https://scryo.loci-labs.ai
The company, not named individual personnel, is responsible for privacy requests. Internal coverage and escalation belong in operating procedures.
Looking for the other document?
Terms of Service