# Ask follow-ups against your saved courses > Your saved lectures become an office hour: ask across all of them and get answers citing the exact minute to rewatch. - For: Student - Category: Learning & Upskilling - MCP tools: deep_context, library_search - HTML version: https://veridive.com/mcp/use-cases/query-your-course-library ## The problem Every studied course produces the same artifact: a pile of watched lectures and a growing list of questions that fall between them. "Didn't she define this differently in week three?" "Which lecture had the worked example with the matrices?" "How does this week's concept connect to the thing from the beginning?" The answers exist — you *watched* them — but retrieving a moment from twenty hours of video is archaeology, so the questions pile up unanswered until the exam makes them urgent. Office hours exist for exactly these questions, and office hours are two hours a week with a queue. What you want is the course itself, answerable: your lectures as something you can ask. ## How veridive MCP helps **deep_context** makes the pile conversational: scope a question to your imported course lectures and it answers from what the lecturer actually said, **citing the exact chunks** — so "where was that worked example?" returns lecture six, minute thirty-four, with the explanation summarized and the source one click away for the rewatch. **library_search** and library_list are the catalog: find the right lectures by title and tag instantly (free), which is how questions get scoped to the relevant weeks. Cross-lecture questions — the connective tissue courses are bad at teaching — become askable: *"how does week eight's method relate to week two's foundations?"* answered across both, cited to both. The course stops being a pile and starts being an office hour with no queue. ## How to run it 1. **Connect veridive MCP** — One-time setup: add the veridive MCP server to Claude, ChatGPT, Cursor or any MCP client and generate your API key. Takes about two minutes — and import your course lectures if you haven't (see the flashcard use case). 2. **Ask the between-lectures questions** — Scoped to the course, answered from the lecturer's own words, cited to the minute. ``` From my imported '' lectures: how did the lecturer define , and did the definition change between the early and later lectures? Cite the moments. ``` 3. **Retrieve the buried moments** — The worked example, the aside, the 'this will be on the exam' — findable by description. ``` Which '' lecture has the worked example about ? Point me to the exact minute. ``` 4. **Build the connective tissue** — Cross-lecture synthesis — the understanding courses assume you'll assemble yourself. ``` Across the '' lectures: how does build on ? Walk me through the chain with citations. ``` ## Problems it solves - Questions that fall between lectures pile up unanswered until exam panic. - Retrieving one moment from twenty hours of watched video is archaeology. - Office hours are two hours a week with a queue; questions don't schedule themselves. - Cross-lecture connections — the real understanding — are left as an exercise for the student. ## Why it's valuable - The course becomes askable: answers from the lecturer's own words, cited to the minute. - Rewatching gets surgical — the cited minute, not the remembered half-hour. - Cross-lecture synthesis on demand builds the connective understanding exams actually test. - The office hour never closes and has no queue. ## Example prompts ``` From my '' lectures: how was defined, and did it change over the weeks? Cited. ``` ``` Which lecture has the worked example on ? Exact minute. ``` ``` How does build on ? Chain with citations. ``` ## FAQ **What needs to be in place first?** The course imported into your library — one playlist import (see the flashcard-deck use case). deep_context then scopes questions to those lectures, up to 20 items per question; library_search finds the right weeks for bigger courses. **Are the answers trustworthy enough to study from?** Every answer cites the exact transcript chunks — the habit is to read the answer, then click the citation for anything exam-critical. The citation is the point: it makes verification a click. **Does this work with the flashcards and cram workflows?** They're one stack: import once, then flashcards for retention, the cram kit for exam nights, and this for the questions in between. Same library underneath all three. ## Related use cases - [Turn a lecture series into a flashcard deck](https://veridive.com/mcp/use-cases/lecture-series-to-flashcards.md) — Import the full playlist and generate spaced-repetition decks per lecture — a semester of review in an afternoon. - [Compress a 3-hour lecture into an exam kit](https://veridive.com/mcp/use-cases/exam-cram-kit.md) — Study notes plus a self-test quiz from one long lecture — study the distilled version, test yourself, go back only where you fail. - [Learn a framework from its creator's own talks](https://veridive.com/mcp/use-cases/interrogate-the-frameworks-creator.md) — Import every talk the creator gave and ask your specific questions — the design rationale docs never capture. --- Part of the veridive MCP use-case library: https://veridive.com/mcp/use-cases.md Server docs for agents: https://veridive.com/mcp.md