oncue

OnCue turns your scattered professional history into perfectly-timed interview answers and writing suggestions. it evolved from the "tellwell" / "storyteller" concept — originally about helping people tell their life stories for college essays, job interviews, networking, and personal branding.

core architecture: multi-modal context is uploaded before an interview (resume, previous interviews, application writing, emails, projects), processed by a model, and during the interview it listens to the interviewer's questions and surfaces relevant experiences to talk about. described as "cluely but context-filled."

memory primitives identified: experience, knowledge, identity, intent, reflection. tech stack: frontend for onboarding and story crafting, backend with LLM APIs, Voyage for vector indexing, and a synthesis layer pointing to evidence for personality traits and patterns.

the market research showed $1.2-2.5B interview prep market growing at 8.9-11.8% annually. differentiation from generic ChatGPT: seamless multi-source context, understanding of user voice through feedback loops, proactive story mining when new things happen. the tellwell pivot from college essays to interview prep came from the insight that "cover letters are actually not that important; more important is the interview."

OnCue became a real project that was presented. connects to the broader me model concept (building a representation of yourself from all data sources) and to the precision description engine for the output quality side.

timeline

• [2025-10-12] initial tellwell concept — storyteller for college/job apps
• [2025-10-16] pivot to interview focus after market research
• [2025-10-18] noted esai competitor only good at asking questions
• [2025-10-19] technical planning — data connectors, local embeddings, app wrapping
• [2025-10-24] sheetal meeting — discussed MVP features, marketing, privacy