V0.9 Conversation-First Signals & Public Research

Signals becomes conversational from creation through ongoing management, while evidence-backed Lucid research becomes accessible before signup.

Conversation-first Signals – Create, refine and activate ongoing monitoring by describing what you want Lucid to watch in natural language.
Signal lifecycle controls – Signals now support clear Draft, Live, Paused and Archived states.
Public research preview – Ask Lucid intelligence questions from the public experience and inspect supporting evidence before creating an account.
Explicit research sources – Select Lucid News and available official datasets as part of the research boundary for a question or Signal.

More relevant Lucid News retrieval – Improved query planning, date handling, publisher diversity and relevance so retrieved coverage better matches the question being asked.
Richer entity context – People, organisations and places can resolve to more useful profiles with identifying and contextual information.
More honest monitoring outcomes – Signals better distinguish between material change, repeated information, quiet periods and insufficient source coverage.
Clearer evidence and provenance – Improved how Lucid communicates which sources were selected and which evidence actually supported a response.
Stronger end-to-end Signals workflow – Creating, activating, running and reviewing a Signal now behaves more like one coherent workflow.

Fixed paused or archived Signals becoming eligible to run.
Fixed cases where a confirmed Signal could fail to save correctly.
Fixed selected sources being omitted from some mixed-source research.
Improved failure handling so unavailable evidence is not confused with evidence that nothing changed.
Fixed reliability issues affecting public research and Resources.