SweepJobs web app, browser extension and API. Last updated 7 October 2026. This describes the current beta implementation and still requires legal and infrastructure-retention review before a broad commercial launch.
SweepJobs helps you search indexed public jobs and prepare applications from information you choose to provide. Your account data is private to your account. The browser extension only operates on supported application pages when you install and use it. You may install the dashboard as a desktop web app through Chrome or Edge; that install still uses the online service. Its service worker caches only a generic offline page and does not cache profile, resume, sign-in, or API responses.
| Data | Why |
|---|---|
| Contact details, work history, education, skills, job preferences, and any work-authorisation or demographic details you choose to enter | Build your profile and prepare application answers and resumes. Sensitive details are reused only when you explicitly enable that setting. |
| Uploaded resume file, extracted text, and generated tailored resume content | Populate your profile, generate resume variants, and attach a resume when an application asks for one. |
| Saved answers, pending questions, application fields and field-resolution records | Show what was filled, remember approved answers, and return unanswered jobs for your review. These records can contain personal answers. |
| Selected jobs, job-run history, watched company boards, filters and notification preferences | Search listings, run only the jobs you select, and report progress. |
| Public ATS board/job URL and timestamp when you explicitly suggest a board, plus an account-linked signal | Deduplicate suggestions and limit abuse. The signal is not shared with other users, appears in your account export, and is removed when you delete your account. |
| Preview summaries and screenshots | Show what a form preview resolved. Screenshots are automatically removed after 30 days; field-level records remain with your account until deletion. |
| Optional quality signals, only if you turn them on in Account settings | Daily counts of broad ATS family, field category/type, resolution outcome, answer source, model when used, and standardized question-catalogue ID. Signals remain linked internally to your account so you can revoke and erase them; they are retained for no more than 90 days. No answer text, raw form label, profile/resume, employer, job title or job URL enters this separate table. The operator report may show the catalogue's standardized wording only for groups with at least 10 contributing accounts. |
| Email address, salted password hash when you set a password, one-time-code digest, requesting IP address, hashed email/IP rate-limit identifiers and signed-in session metadata | Verify password or email-code sign-in, limit abuse, and keep your account signed in. We do not store plaintext passwords or sign-in codes. |
| Optional GitHub username and public repository metadata | Only when you choose to sync public projects from GitHub. |
Some fields are never inferred by the model: legal attestations and certifications, consent checkboxes, electronic signatures, citizenship, criminal history, salary expectations, government identifiers, and the contact details of your references. You may enter your own work-authorisation, sponsorship, or optional demographic facts in your profile and explicitly enable sensitive autofill. Those values are reused only when they match the choices on the form; otherwise the field stays blank for you.
Account data is stored and processed on Cloudflare services, including D1, R2,
Workers, Workers AI and Browser Rendering. The current Worker configuration uses
@cf/google/gemma-4-26b-a4b-it through Cloudflare Workers AI. Relevant
profile, answer and job text may be sent to that configured model for generation. The
service does not operate a separate model-training pipeline using customer profiles.
Sign-in and optional run-completion emails use Cloudflare Email Sending. If you ask to sync GitHub, the service queries GitHub's public API for public repository metadata. Job discovery fetches public listings from the selected Greenhouse, Lever or Ashby company-board APIs. If autonomous web discovery is enabled, fixed ATS-focused discovery queries and public result URLs are sent to the Brave Search API; user identities, resumes, profiles, and personal job-search terms are not included. No payment processor is active; no payment information is collected.
We do not sell account data or use it for advertising. Account profiles, answers and resumes are not shared with other SweepJobs users. The shared question catalogue stores question wording and categories, not your answer values. When you submit an application, the information in the form is sent to that employer's applicant-tracking system by the browser, as it would be if you typed it yourself.
If you opt in to quality signals, the service may use them for internal product improvement. The account link is retained only to enforce your controls and deletion; the operator report contains grouped counts only and suppresses every group with fewer than 10 distinct contributing accounts. No account identifier is returned by that report, and no third-party analytics provider receives these signals. The signals help identify product gaps; they do not train or fine-tune a model.
Applications are filled in your own browser. By default, the extension does not submit: you review and submit yourself. Optional unattended submission is off by default and stored only in the browser extension's local settings; no server setting can turn it on. A form the extension cannot safely complete is left for review.
The software keeps a shared catalogue of the questions job applications ask — the wording employers use, and which phrasings mean the same thing. That catalogue improves for everyone as the software is used. It contains no personal data: questions are shared, answers are not. Your answers, corrections, resume and profile are not shared with other SweepJobs users; the service-provider processing and employer transfers described above still apply.
The list of company job boards being watched is also shared, so a board is checked once rather than once per person. Those are public job listings. If you explicitly suggest a board, only the supported ATS board and its public postings may enter that shared catalogue after verification; your account-linked suggestion signal remains private.
Account deletion covers active application data. Provider-managed backup and restore retention settings still need to be verified and described before broad commercial launch; this beta notice is not a claim that every backup copy disappears immediately.
Questions or privacy requests: tyler.gee13@icloud.com
See also the Terms of Service and Help & support.