Transparency Report

Third Eye Cyborg, LLC publishes this Transparency Report on a semi-annual basis covering two fixed reporting periods each calendar year: January 1 – June 30 and July 1 – December 31. Each edition is published within 90 days of the close of the period it covers.

The first edition under this cadence will cover the reporting period July 1, 2026 – December 31, 2026 and will be published in final form on or before March 31, 2027. The categories, methodology, escalation paths, and contact channels described in the sections below are in effect now and apply to every legal-process item, takedown notice, government request, and moderation action recorded from the publication date of this page forward, so that the first edition can be built directly from contemporaneous records.

Loading report data…

Counts in this report are aggregated from our internal compliance ledger, which records every legal-process item received by Third Eye Cyborg, LLC and every moderation action taken under our Acceptable Use Policy, DMCA Policy, DSA Notice and Action mechanism, and Child-Safety workflow. Each entry is timestamped, classified by request type, and reconciled against the originating record (notice, warrant, legal request identifier, or in-product report ID) before it is included.

Where a single underlying event generates multiple records (for example, a single warrant covering multiple accounts, or a single DMCA notice covering multiple URLs), we count one request per underlying instrument and one action per affected account or URL, and we identify that distinction in the relevant subsection.

Automated accounts, including Network Bots, are a tracked category in the compliance ledger. Bot-related notices, content actions, token revocations, and installation removals are aggregated with the relevant moderation or enforcement category and identified separately where reporting volume permits without compromising privacy.

The categories below cover the July 1, 2026 – December 31, 2026 reporting period. Sub-jurisdiction breakdowns are provided where the number of requests in a jurisdiction is large enough that doing so does not compromise the privacy of users or the integrity of ongoing investigations.

• US subpoenas received and number complied with in whole or in part .
• US court orders under 18 U.S.C. § 2703(d) received and complied with .
• US search warrants received and complied with .
• US National Security Letters and FISA process received , reported within the bands permitted by 50 U.S.C. § 1874.
• Emergency disclosure requests under 18 U.S.C. § 2702(b)(8) received and acted on .
• Preservation requests under 18 U.S.C. § 2703(f) received and honoured .
• Non-US government requests received , broken down by jurisdiction where volume permits.
• Number of users / accounts identified across all requests above: .
• Percentage of requests that produced some user data: .

• 17 U.S.C. § 512(c) takedown notices received .
• Notices acted on in whole or in part , and notices rejected as facially invalid .
• Counter-notices received under § 512(g) and content restored .
• Repeat-infringer terminations carried out under our published repeat-infringer policy .

• Article 16 notices of illegal content received from recipients of the Service .
• Notices received from trusted flaggers under Article 22 .
• Statements of Reasons issued under Article 17, broken down by action type (content removal, visibility restriction, monetisation restriction, account suspension, account termination) .
• Internal complaints lodged under Article 20 and their outcomes — upheld , reversed , partially reversed .
• Out-of-court dispute settlement proceedings under Article 21 in which we participated .
• Median and 90th-percentile time to action for each category: .
• Share of decisions in which automated systems were used: .

• Child-safety / CSAE reports received through in-product reporting .
• The current in-product workflow is authenticated, rate-limited, deduplicated, and queue-first. Child-safety report descriptions are omitted from the ordinary report record; only the category, target identifiers, and limited routing metadata are retained for review and accountability.
• NCMEC/CyberTipline reports, automated CSAM detection or hash matches, automated screening for every report, evidence preservation, and authority referrals are not represented as operational in this report until the responsible operator confirms those processes.
• Accounts terminated for child-safety violations are reported only when a reviewer records that outcome.

• Accounts actioned for Acceptable Use Policy violations, broken down by the most-cited violation categories .
• Posts, books, chapters, Library videos, and other content units removed or restricted .
• Appeals received and the share that were granted in whole or in part: .

Appeals marked as auto-denied represent submissions that reached their review due date without an operator decision in the configured SLA window.

This section also includes enforcement actions involving automated accounts and Network Bots where policy-violating automation impacted users, channels, or network governance.

We design our systems so that we do not retain data we do not need. As a result, certain figures other platforms publish are not meaningful for Ascended Social — for example, we do not run third-party behavioural advertising at all, so we have nothing to report under "ad-targeting categories shared". We do not sell personal information for money and do not share personal information for cross-context behavioural advertising as defined by the CCPA / CPRA, so the corresponding "sale" and "share" figures are structurally zero.

For the legal framework that drives these figures, see the Law Enforcement Guidelines, the DMCA / Copyright Policy, the EU DSA Notices, the Child-Safety Statement, and the Acceptable Use Policy.

Questions about this report can be sent to [email protected]. Researchers and civil-society organisations who would like the underlying methodology spreadsheet for a published period may request it from the same address.