You applied to twelve roles in the past four weeks, three of them have come back with phone screens, and you genuinely cannot remember which version of your resume you sent to the company you're interviewing with on Thursday. The notes on the role are in a saved tab, the recruiter's email thread is somewhere, the company's product is in a different tab, and the interview is in forty-eight hours. You sit down to prep, spend the first twenty minutes just reassembling context, and then realize you only have an hour for the actual prep.

Most people in an active job search run into this version of the problem. The information about every application — the role, the company, the people, the resume version, the conversation history, the interview rounds — lives across LinkedIn, email, ATS portals, browser tabs, and your fading memory. By the time you actually need to walk into an interview prepared, you're spending your prep time reassembling instead of preparing.

A vault that holds one parent page per application, with the role, the company research, the interview notes, and the follow-ups, fixes most of it. The agent does the assembling when something comes up.

## One application page, one company underneath

In Docapybara, each role you're actively pursuing gets a parent page. Pages nest with no depth limit, so each application can have child pages for *Role* (job description, salary band, recruiter contact), *Company* (research, product, news, people), *Application* (the resume version sent, the cover letter, the date submitted), *Interviews* (one child per round), and *Follow-ups*.

If you're tracking a high volume of applications, group them under an *Active applications* parent page, with a separate *Past applications* page for ones that have closed. The agent searches across when you need the cross-cut. *"Which roles in the past three months had hybrid options and a salary band above $X?"* The answer comes back grounded in the actual application pages.

For the broader career-transition shape — including the question of whether to leave at all — see [Career Transitions and Job Searches, With a Vault Behind You](/guides/personal-life/career-transitions-job-search/). This guide focuses on the active-search mechanics.

## A pipeline view that keeps the search legible

The pipeline is the workhorse during an active search. An inline database via the `:::database:::` directive on the *Active applications* page handles it. Columns for company, role, status (applied, screen scheduled, screen done, onsite, offer, declined, ghosted), application date, last contact date, salary band, and notes.

The agent can update this from voice or text. *"Update the [Company] application — phone screen done yesterday, moving to onsite next week."* The row changes. *"Show me everything in the pipeline that hasn't moved in two weeks."* You get the list of roles to follow up on or close out.

The reason this matters more than a spreadsheet is the cross-cut analysis. *"What's the average time between application and first contact for roles in this pipeline?"* *"Of the screens I've done, what's the conversion to onsite?"* The agent reads across the database and answers, which is hard to do from a static spreadsheet without doing the math yourself.

Pipeline visibility also helps the small psychological battle of a long search — looking at twelve active applications instead of one rejection makes the situation more legible.

## Company research that actually informs the interview

The company page is where the prep work pays off. The role's job description goes here. Recent company news — funding rounds, product launches, leadership changes — drops on the page. The product itself, if you can try it, gets a few notes. The names of the people you'll be interviewing with go in a *People* section.

For company research, the agent's `web_search` tool can pull current information. *"Find the most recent news about [Company] in the past six months — funding, product launches, leadership changes."* The summary comes back with sources. Save what's relevant on the page.

For the people you'll be talking to, LinkedIn profiles, recent posts, podcast appearances, blog writing — anything publicly visible — informs how the conversation goes. *"For [interviewer name], pull what's publicly visible about their background and any recent writing or interviews they've done."* The summary lands; you skim before the call.

This isn't about being a stalker — it's about walking into the conversation with enough context to ask real questions back. The interviewers can usually tell when you've done the homework. The hour you spend on this is worth more than the hour you'd spend re-reading the job description for the fifth time.

## The resume and cover letter you actually sent

Track which version of your resume you sent to which application. This sounds trivial until you're tailoring versions for different job families and three weeks later you can't remember which one a specific company saw.

A *Resume versions* parent page holds each variant — the generalist version, the IC-leaning version, the manager-track version, the contract version, anything else you maintain. Each application page references which version was sent. The agent can answer *"What resume did I send to [Company]?"* with the answer and a link to that version.

Same for cover letters and any portfolio links. The artifact you submitted is part of the application's record, not just a file in a downloads folder. Three weeks later when the interviewer references something in your cover letter, you know what they're looking at.

For longer-form applications — academic positions, government roles, fellowships, anything with a multi-document submission — the agent can help draft and refine. *"Draft a cover letter for [role] using [resume version] as context, focused on the parts of the job description that match."* You get a draft to refine; the role's page already has everything the draft needs.

## Interview notes that survive the round

Each interview round gets its own child page on the application. *Phone screen with [name] — [date]*, *Onsite round 1: [name] — [date]*, etc. The page holds the prep notes (questions you want to ask, points you want to make, weaknesses to anticipate) and the post-interview notes.

For the interview itself, voice notes immediately after capture what was actually asked and discussed. *"Phone screen with [name], 30 minutes. Walked through my background, then they asked about a project where I'd had to influence without authority. Spent ten minutes on a hypothetical about [scenario]. They mentioned [team detail] in passing. Next step is a take-home, due Friday."* The transcript drops on the round's page.

Two days later when you're prepping for the next round, the agent can pull what you learned. *"Summarize what we know from the previous rounds with [Company]: who I've talked to, what they asked, anything they mentioned about the team or the role."* The summary lands grounded in the actual round notes. You walk into the next round with continuity instead of starting from zero.

For more on the broader voice habit, see [The Complete Guide to Voice-First Note-Taking](/guides/personal-life/complete-guide-voice-first/) — the post-interview download is a textbook example of where voice beats typing.

**Take-homes, case studies, and the assignment work.** For roles that include a take-home assignment or case study, the assignment lives on its own child page. The brief goes there, the work product goes there, any clarifying email exchanges go there. For technical assignments specifically, the agent can answer questions about your own past work as you draft — *"Pull from my past project notes any examples of [pattern] I've worked on."* If you've kept your own work log, the answer comes back grounded in it.

## Follow-ups, references, and the offer stage

After every meaningful contact — phone screen, onsite, conversation with a recruiter, networking call that touched on a role — capture the follow-up that's owed. *"Send a thank-you note to [interviewer] referencing the [topic] we discussed."* The agent can draft it from the round's notes. You review and send.

For references, a *References* page holds the people who've agreed to be references, with the kind of role they're best positioned to speak to and any prep notes you've shared with them. When a company asks for references, the agent can pull the right list. *"Who would be the strongest references for an IC engineering role?"* The answer comes back from the page.

At the offer stage, an *Offers* page holds each offer in writing — the comp breakdown, the timeline, the negotiation history. For comparing multiple offers, the agent can lay out the trade-offs. *"Compare the offers from [Company A] and [Company B] across base, equity, vesting, sign-on, benefits, and trajectory."* The summary comes back grounded in the actual offer letters.

For the broader version of weighing options, see [Capture and Compare Options for Any Major Decision](/guides/personal-life/capture-compare-any-decision/). A job offer is one of the higher-stakes versions of that pattern.

## The post-search archive

When the search closes — accepted offer, declined offer, paused — the active pages move to a *Past applications* parent. The notes don't disappear; they become reference material for the next time.

A year or two later, when you're considering a new search, the past pages are gold. *"Which companies last year made it to onsite? Which roles did I find energizing in the prep stage?"* The patterns inform where to look first.

## A starter shape for an active search

If you're starting a job search next week:

- **One pipeline database** on the *Active applications* page.
- **One parent page per application** as it goes from interest to active.
- **A company child page** with research, news, and people.
- **An interviews parent** with one child per round.
- **A resume versions page** referenced by each application.
- **A follow-ups view** the agent can pull on demand.

That's it. No template, no taxonomy beyond company name and round.

The point isn't to make the job search itself into a record-keeping project. It's that the small amount of structure you keep means each interview starts from real continuity, the prep time goes to actual prep, and the patterns of what's working in the search become visible enough to act on.

[Try Docapybara free](/accounts/signup/) — start with the application that's at the top of your mind today, and the next interview you walk into, the prep will be on the page waiting for you.