Transparency & Accountability Report
Our commitment to openness about how Parsafy operates, moderates content, and responds to legal and government requests.
At Parsafy AI, transparency is not a compliance exercise — it is a founding commitment. Our users, our community, regulators, researchers, and civil society deserve to understand how Parsafy operates: how we moderate content, how we respond to law enforcement, how our advertising systems work, and how we are performing against the safety commitments we have made.
Parsafy has not yet launched publicly. Our inaugural Transparency Report will be published following Parsafy's official platform launch and will cover the initial operating period. In the meantime, this page describes what our Transparency Reports will contain, our reporting methodology, and our specific commitments under applicable transparency regulations including the EU Digital Services Act (DSA).
We invite researchers, journalists, regulators, and civil society organizations to hold us accountable to the commitments described here. Transparency is only meaningful when it is specific, measurable, and published on a regular schedule — and that is what Parsafy commits to.
📋 Our Transparency Commitments
Parsafy commits to publishing Transparency Reports on a regular cadence following platform launch. These reports will be published at parsafy.com/transparency-report/ and will be freely accessible to all users and the public without any login requirement.
Our Transparency Reports will be specific and data-driven — not vague summaries. Every category will include actual numbers: reports received, actions taken, appeals filed and overturned, law enforcement requests received and responded to. We will report on what we got wrong as well as what we got right.
We will comply with all applicable transparency reporting requirements, including those under the EU Digital Services Act, GDPR, and California's transparency requirements under the CCPA and AB 587.
Child Sexual Exploitation & Abuse (CSEA)
Volume of CSAM reports received, confirmed cases, NCMEC CyberTipline reports filed, accounts terminated, and law enforcement referrals. Detection rates comparing proactive automated detection vs. user-reported detection.
⏳ Post-LaunchTerrorist & Violent Extremist Content
Volume of terrorism and violent extremism reports received, content removed, accounts terminated, and proactive detection rates. Breakdown by category (recruitment, propaganda, glorification). Law enforcement referrals.
⏳ Post-LaunchSelf-Harm, Suicide & Eating Disorders
Volume of self-harm and crisis content reports received. Actions taken — including content removal, resource surfacing, and emergency referrals. Proactive detection rates and crisis resource activation rates.
⏳ Post-LaunchHarassment, Hate Speech & Threats
Volume of harassment, hate speech, and threat reports received by subcategory. Content removed, accounts actioned, and enforcement actions by severity tier. Community Notes activity on contested content.
⏳ Post-LaunchMisinformation & Deceptive Content
Volume of misinformation and impersonation reports received. Content removed, Community Notes added, and coordinated inauthentic behavior actions. Electoral integrity enforcement during election periods.
⏳ Post-LaunchIllegal & Regulated Activities
Volume of illegal activity reports by subcategory (drug sales, trafficking, fraud, cybercrime, weapons). Enforcement actions taken and law enforcement referrals. Breakdown by violation category.
⏳ Post-LaunchContent & Account Appeals
Total appeals received by type (content vs. account). Appeal outcomes — upheld, denied, partially upheld. Overturn rate by violation category. Average time to appeal decision. EU out-of-court dispute settlement referrals.
⏳ Post-LaunchAdvertising Moderation
Volume of ads reviewed, rejected, and removed. Rejection rates by category (political, prohibited, restricted). Advertiser application approvals and denials by category. Ad-related user reports received and acted on.
⏳ Post-LaunchLaw Enforcement & Government Requests
Volume of legal process received by type (subpoena, court order, search warrant) and by country. Response rates, preservation requests complied with, and emergency disclosures made. Government content removal requests received and compliance rate.
⏳ Post-LaunchCountry & Regional Overview
Country-level breakdown of reports received, content actioned, law enforcement requests, and active users. EU member state data as required by the Digital Services Act, including monthly active user counts and content moderator allocations.
⏳ Post-LaunchAI Tools & Automated Moderation
Proactive vs. reactive detection rates across violation categories. AI false positive rates and correction rates following human review. AI-generated content moderation performance. Hash-matching detection statistics.
⏳ Post-LaunchCommunity Notes
Total Community Notes submitted, approved, and displayed. Notes by category (health, politics, misinformation, other). Notes that led to content removal vs. context-only notes. Contributor statistics and quality ratings.
⏳ Post-Launch🇪🇺 EU Digital Services Act (DSA) Transparency Obligations
Parsafy AI Inc. acknowledges and commits to complying with its transparency reporting obligations under the EU Digital Services Act (DSA). Our Transparency Reports will include DSA-specific disclosures as required, including:
- Number of monthly active users (MAUs) in the European Union, by member state where available.
- Number of content moderators allocated to EU-facing moderation, broken down by language.
- Volume of reports of illegal content received under DSA Article 16, by category and member state.
- Volume and outcomes of DSA Article 17 notices of illegal content submitted by trusted flaggers.
- Volume and outcomes of appeals submitted under DSA Article 20, including out-of-court dispute settlement referrals under DSA Article 21.
- Government and law enforcement requests received from EU member states, by country and request type.
- Any use of automated decision-making systems in content moderation, including accuracy and error rates.
Parsafy will designate an EU legal representative and comply with all applicable DSA obligations commensurate with our platform size and category. DSA transparency reports will be submitted to the relevant Digital Services Coordinator as required.
Publication Schedule
Parsafy will publish Transparency Reports on a semi-annual basis — covering January through June and July through December of each year. The inaugural report will cover the period from Parsafy's public launch date through the end of the first complete six-month reporting period. Reports will be published within 90 days of the close of each reporting period.
Data Standards
All statistics will be reported as actual numbers where possible. Where rounding or minimum thresholds apply (to protect individual user privacy), we will clearly note the methodology used. We will not report vanity metrics without context — every number will include the denominator or base rate needed to evaluate it meaningfully.
Proactive vs. Reactive Detection
For all content moderation categories, we will distinguish between content identified proactively by our automated systems or Trust & Safety team, and content identified reactively through user reports. This distinction helps researchers and policymakers understand the true effectiveness of our detection systems.
Geographic Breakdown
We will provide country-level data for all categories where volume is sufficient to be reported without compromising individual user privacy. For EU member states, we will provide the additional detail required by the DSA. For countries with very low report volumes, data may be aggregated into regional groupings.
Third-Party Verification
Parsafy is committed to exploring independent third-party verification of our Transparency Report data as our platform matures. We will also make data available to qualified academic researchers studying online safety and platform governance, consistent with user privacy obligations.
Transparency is a commitment, not a checkbox. We look forward to being held accountable to it.
Safety Policies · Community Guidelines · Moderation & Appeals · Law Enforcement