AI Transparency & Content Policy

Last updated: August 11, 2026

We want you to know exactly what you are working with. Jervis is Plexa AI's assistant: an orchestrated system with its own memory, skills and identity layer that plans work, retrieves live information, and produces answers, media and automation.

Capabilities

Conversational reasoning, research with live web retrieval and citations, long-context recall of your workspace memories.
Autonomous multi-step agents that can plan, execute and report on tasks you approve.
Media generation: images, video, thumbnails and voice-assisted input.
Publishing and SEO tooling with scoring, hooks and competitive comparison.

Known limitations

Output can be wrong, incomplete, out of date or confidently mistaken ("hallucination").
Retrieval depends on what is publicly reachable at request time; sources can themselves be inaccurate.
Numerical, legal, medical and financial reasoning must be independently verified.
Similar prompts can produce different results; generation is not deterministic.

Human oversight

You remain the decision-maker. Do not deploy Output in consequential contexts without qualified human review. Agents ask for approval before destructive or external-facing actions, and every run is logged so you can audit what happened.

Provenance and labelling

Generated media is intended to be labelled as AI-generated when published. Do not remove or obscure provenance markers, and do not present generated content as an authentic recording of real events.

Safety systems

Requests are screened against our Acceptable Use Policy. Safety filters can produce false positives; if you believe a legitimate request was blocked, contact support@plexaai.app with the details and we will review it.