Frequently asked questions.

A self-hosted job-search assistant. It runs nightly scans, ranks new postings against your resume, and on-demand generates a tailored resume, cover letter, fit analysis, company brief, and interview prep for any role you choose to pursue.

The platform itself is free for users on a friends-only deployment. The admin running the deployment pays for the LLM tokens against an Anthropic key; per-user daily caps prevent runaway spend.

Click Sign up in the top nav, enter your email + password, then click the verification link the admin sends you.

Beta note: EasyApply is currently in private beta. New signups are gated by manual admin approval — the admin reviews your sign-up and may contact you directly before issuing the verification link.

Add your base resume (Base resume tab). PDF upload, paste markdown, or upload a .txt — whichever is fastest. The rest of the system is downstream of that one document.

One saved nightly query. It bundles a search term, location, job-type filter, freshness window, and any keyword exclusions. The scan engine runs each active profile once per night against the enabled job sites and dedups results per user.

On the Search profiles page, click Recommend search profiles. The LLM reads your resume and proposes 3–5 drafts (created inactive so you can review them first). Edit names / locations as needed, then flip the Active toggle.

One of three reasons: (a) you don't have a base resume yet — add one first; (b) you're at your max_search_profiles quota — delete one to make room; (c) you've used your daily LLM-document budget — try again tomorrow.

Each active search profile fires once per night, jittered into the 03:00–07:00 UTC window. The same profile lands at the same minute every night so the spread is stable; different profiles spread across the window to avoid hammering the upstream sites.

Open the Scans tab and inspect the most recent run. Common causes: zero scan sites enabled (check the Profile page), search term too narrow, location too specific (try state-level first), or the upstream board rate-limited the deployment. The run page shows the error message if there was one.

The current status is mutable — you can move it forward or backward when you correct mistakes. The highest rung climbed is monotonic and never decreases. Funnel analytics use the monotonic value so a back-edit (e.g. you accidentally clicked "Submitted" too early) doesn't lie about your application history.

Each new job from the nightly scan goes to Sonnet alongside your base resume + that profile's intent. The model returns a tier (S/A/B/C) and a one-paragraph rationale. S = clear yes, C = skip.

The system prompt explicitly forbids invention of titles, dates, or metrics. It can reorder + reword bullets to match the JD's emphasis, but it can't add experience you don't have. That said: read the output. The LLM is a draft assistant, not a final author.

Yes — every generated document has a markdown editor on the application detail page. Edit, save, then click the PDF download for a fresh render. Editing is preferred over regenerating because regeneration costs tokens and may rewrite parts you liked.

It runs Sonnet's web_search tool to gather current public info (recent news, funding, leadership, strategic moves) rather than relying on training-data knowledge. A few search iterations + synthesis + emit is unavoidably slower than a plain LLM call.

Your account has a max_documents_per_day cap (default 10) shown on your Profile page. Tailored resumes, cover letters, fit analyses, interview preps, company profiles, and search-profile recommendations all count toward it. Resets at UTC midnight.

Yes. On Profile → PDF configuration, click Download .css on any built-in template (Classic, Modern, Compact), edit the CSS, and upload it back into the Custom slot. CSS is the only customisable layer; the HTML wrapper stays server-controlled for security.

On the Application detail page, the Resume and Cover-letter tabs each have a Download PDF button next to the editor. PDFs are re-rendered on every click using your currently-selected template.

On the server hosting your EasyApply deployment. Resume, applications, and generated documents live in Postgres + on-disk files under storage/users/<your-id>/. No third-party storage. No sharing across users.

No. The admin dashboard exposes aggregate counts and per-user quotas only. There is no admin surface that reads document contents.

The relevant subset for whatever it's doing: your base resume (always), the JD (when tailoring/fit/prep), your existing search profiles (when recommending), and the company name + role (when producing a company brief). LLM calls go to Anthropic with the deployment's system key.

Account-deletion UI is post-v1; for now, ask the admin. Deletion wipes your base resume, applications, generated documents, conversation history, search profiles, and usage logs.