Scrappy Singh

Scrappy Singh is a personal AI operating system created by Karnveer Singh. It exists to keep one consistent reasoning partner across every AI model he uses, instead of restarting context in a new chat window every day.

It behaves as a mentor and collaborator: it remembers long-term context, recognises patterns across conversations, and answers in a personalised reasoning style rather than a generic assistant tone.

A prompt boot system loads identity, memory and reasoning rules at the start of any session, which is what makes the same personality portable across models.

A long-term memory layer stores durable facts, decisions and open threads separately from short-term conversation context, so the system can evolve its understanding over months.

Observation Mode runs alongside normal answering and surfaces things that were never asked: contradictions, missing information, hidden opportunities and unusual patterns.

Persona switching selects the right reasoning mode — engineer, researcher, strategist — based on the task instead of the phrasing.

Keeping memory useful instead of noisy: most stored context degrades answers, so retrieval is scoped and decayed rather than dumped into the prompt.

Model portability: each model interprets system instructions differently, so the boot prompt is written to be behaviourally, not syntactically, identical.

Making proactive observation feel helpful rather than intrusive.

Scrappy Singh is the working laboratory behind Karnveer's research into AI memory, pattern recognition and human-AI collaboration.