A research-support product for people who are early in a project and need structured direction, not just another chat window. It remembers project context, tracks research stage, and ties its suggestions back to real sources.
Context
AI Research Supervisor is a research support product I’m building for people (especially beginners) who are early in a research project and need structured help choosing a direction, not just a chat window.
Problem
Most AI research and writing tools behave like generic chatbots: no memory of the project between sessions, no sense of what stage the research is actually at, and no clear link between a suggestion and the evidence behind it. For someone new to research, that makes the output hard to trust, or even use.
Product Approach
Instead of a single-shot chat interface, the product is built around three linked ideas: it keeps a persistent record of the project, reads that against the current research stage, and only then produces a recommendation it can trace back to a source.
Project Context
Remembers what the user is working on across sessions, so nobody starts from zero each time.
Research Stage
Reads whether the user is choosing a theory, reviewing literature, or further along, and adapts its response accordingly.
Evidence-Aware Recommendation
Produces a suggestion designed to be traceable back to a real, identifiable source rather than open-ended generation.
Current Capabilities
Current Limitations
An honest list of what still needs work:
What I’m Improving Next
The immediate priorities are fixing the upload limits, smoothing out the chat flow during theory selection, improving the consistency of paper recommendations, and resolving the interaction bugs above.
