Roadmap
Disclosure, content labeling, and audit logging are patterns, not one-off features — Disclarion is built so that meeting the next regulatory deadline is a config change, not a rewrite.
Below is the actual EU AI Act calendar driving this roadmap, what we've already shipped, and what's next, later, and still exploratory — in that order, so it's clear which is which.
Four dates. Everything else on this page maps to one of them.
Article 50 transparency obligations apply
How Disclarion covers this: Chatbot disclosure, content labeling, and audit-ready logs.
Art. 50(2) machine-readable-marking grace period ends for generative systems already on the market before 2 Aug 2026
How Disclarion covers this: Disclarion adds Content Credentials.
High-risk obligations for Annex III systems apply (deferred from 2 Aug 2026)
How Disclarion covers this: High-Risk Readiness evidence layer.
High-risk obligations for Annex I systems apply — AI embedded in regulated products (deferred from 2 Aug 2027)
How Disclarion covers this: Exploring fit with sector conformity workflows.
Article 50 violations carry penalties up to €15 million or 3% of worldwide annual turnover, whichever is higher. For SMEs and start-ups, the AI Act applies the lower of the two amounts (Art. 99(6)).
Shipped, in production, covering Article 50 as it stands today.
Chatbot disclosure, content labeling, and audit-ready logs
The core loop: the SDK flags when a disclosure is owed on every AI interaction, every generated output is labeled at the source, and every record is written to an insert-only, exportable trail. For any team running OpenAI, Anthropic, or Gemini under EU AI Act Article 50.
Labeling beyond chat
The same disclosure and audit infrastructure covers non-chat AI-generated content too: articles and product descriptions get a positioned HTML label, and generated images get a real watermark baked into the file itself via track_content().
Public record verification pages
Every "Labeled by Disclarion" credit links to a public /verify/{id} page confirming that exact record is real — when it was logged, provider, content type, disclosure status. Metadata only, never the content itself, and never a claim about the content's accuracy.
REST API for any backend
Every SDK capability, disclosure, content labeling, and audit logging, is also available as a plain REST API, for teams on any language, not only Python. The upcoming TypeScript/JavaScript SDK wraps this same API in typed, idiomatic calls; it isn't the only way to integrate today.
n8n community node
n8n-nodes-disclarion brings disclosure, labeling, and audit logging to n8n workflows without writing SDK code directly.
Free public transparency scanner
Point it at any URL for an instant read on which AI-disclosure signals are, and aren't, present. This is a heuristic pre-check, not a certified compliance audit.
Coverage-gap dashboard and one-click evidence reports
A coverage-gap view shows which jurisdiction and interaction-type combinations are logged with no matching obligation rule. A single click turns a date range into a standalone PDF compliance evidence report.
Ahead of the 2 Dec 2026 machine-readable-marking grace period ending.
Content Credentials
Signed provenance for everything Disclarion labels, not just a visible badge.
Continuous Compliance Monitoring
The scanner stops being a one-time check and starts watching your surface for you.
TypeScript / JavaScript SDK
A typed, idiomatic wrapper around the REST API Node teams can already call directly today — same disclosure, labeling, and audit trail, without hand-rolling HTTP calls.
Ahead of Annex III high-risk obligations applying 2 Dec 2027.
Tamper-Evident Audit Trail + Auditor Room
Turn the existing audit log into something a third party can verify without trusting Disclarion's word for it.
High-Risk Readiness
Evidence and documentation tooling ahead of the 2 Dec 2027 Annex III deadline — not a conformity assessment or certification.
Record-keeping here builds on the tamper-evident audit trail above, applied to Annex III's specific documentation requirements, not a second, separate log. This produces evidence and documentation to support a conformity assessment your team or a notified body still has to perform — Disclarion doesn't certify or assess conformity itself.
Directions we're investigating, not commitments — including the 2028 Annex I horizon.
Voice & Video Agent Disclosure
The same disclosure pattern chat interfaces get today, for voice and video agents.
Regulation Radar
Know when a rule you rely on changes, before it becomes a gap in your coverage.
Everywhere you build
The same disclosure and audit infrastructure, wherever the AI call happens.
Annex I compatibility
Exploring fit with sector conformity workflows ahead of the 2028 Annex I horizon.
What actually moved from this roadmap into production. Full, dated history on the changelog page.
Published: disclosing AI on a phone call
Added Article 50 reference and compliance-guide pages
SDK v0.4: real image watermarks and positioned HTML labels
SDK v0.3.2: track_content() for non-chat AI-generated content
Coverage-gap view and evidence reports moved to Now
SDK v0.3: jurisdiction-aware Obligation Engine
Blog launched with the Article 50 explainer
Public AI transparency scanner shipped
Most AI-compliance products stop at a frontend disclosure banner. Ours starts with an audit trail tied to the real LLM call itself, across OpenAI, Anthropic, and Gemini, with real content labeling instead of a static badge. That evidence infrastructure, not the banner sitting on top of it, is what regulators and legal teams actually check for, and it's why each item above extends the same audit trail instead of bolting on a separate system.
Disclarion never stores your conversation or content text — only the compliance metadata needed to demonstrate disclosure and labeling happened.
Roadmap items are plans, not commitments, and are not legal advice. Regulatory dates reflect our reading of the EU AI Act as amended by the Digital Omnibus; confirm with your counsel.
This is an AI-powered assistant.
