You are interacting with AI
Nexpot responses, pointers and Autopilot action plans are generated by artificial intelligence. They are not a human operator. The active Autopilot takeover frame and status capsule identify when AI is using the computer.
Depending on the selected feature, Nexpot uses hosted language, vision, speech-recognition and speech-generation models supplied by Anthropic, OpenAI, AssemblyAI and ElevenLabs. Provider routing can change for reliability and model configuration.
Inputs and outputs
For a user-invoked request, the model may receive the typed or spoken instruction, recent conversation context and full visible-screen images. Voice mode sends microphone audio for transcription. Autopilot receives the task and screen images and returns a proposed next action such as click, type, select, scroll, wait or navigate.
The system can misunderstand the screen, point to the wrong place, type incorrect content or behave unexpectedly when interfaces move. Output is probabilistic and must be checked against the visible result.
Human control
Nexpot does not make decisions producing legal or similarly significant effects about people. It is not designed to replace professional medical, legal, financial, employment or safety judgment.
Safety and crisis response
Nexpot applies provider safeguards and a local high-confidence crisis response for direct self-harm or suicide intent. That response encourages immediate human help and does not consume credits. Nexpot cannot monitor a person, contact emergency services or guarantee detection.
If there is immediate danger in the EU, call 112. In Hungary, the free 24/7 emotional support line is 116-123. In the United States and territories, call or text 988. Elsewhere, contact local emergency services or a trusted person nearby.
Data and questions
Raw screenshots and microphone audio are not intentionally stored in Nexpot's application database after the request, but providers may retain request data temporarily under their terms. See the Privacy Notice for purposes, legal bases, retention and rights.
Report a harmful or incorrect result to hello@nexpot.ai. Include only the minimum screenshot or context needed and remove unrelated sensitive information.