Running your first sweep
You've created your configuration profile. Time to run your first automated job sweep.
Run the sweep
Open your terminal in the folder that holds your sweepy clone. If you closed the terminal since the last step, reopen it and go back to sweepy:
cd sweepy
Run the engine with your profile:
sweepy --profile profiles/paralegal_nyc.yaml
Sweepy includes paralegal_nyc.yaml in the profiles folder. The short form also works:
sweepy --run
--run reuses the last profile you saved.
If you built your profile with the wizard, Sweepy saved it as profiles/alex_rivera_paralegal.yaml. The name combines your name and your first job title. Use that path instead of the bundled preset.
Understanding the terminal output
Sweepy prints real-time progress messages as it works. You don't need to interact with it. Just watch and wait.
The terminal output below comes from Alex's paralegal sweep. Treat it as illustrative: exact counts, filenames, and scores change with every run.
================================================================================
Sweepy v3.0.0
Profile: Alex Rivera — Paralegal, Legal Assistant, Litigation Support Specialist, Legal Secretary (New York, NY)
Execution method: cli
================================================================================
[1/5] Loading profile...
* Profile loaded: profiles/paralegal_nyc.yaml
* Search titles: Paralegal, Legal Assistant, Litigation Support Specialist, Legal Secretary
* Location: New York, NY (including remote)
[2/5] Discovering roles...
* Cross-board duplicate checking: enabled
-> Searching Greenhouse............... Found 0 postings
-> Searching Ashby.................... Found 1 postings
-> Searching Lever.................... Found 0 postings
-> Searching Workable................. Found 1 postings
-> Searching SmartRecruiters.......... Found 0 postings
-> Searching Breezy HR................ Found 0 postings
-> Searching Glassdoor................ Found 30 postings
-> Searching Indeed................... Found 20 postings
-> Searching LinkedIn................. Found 66 postings
-> Searching Built In NYC............. Found 19 postings
* Total discovered: 137 postings
[3/5] Validating URLs...
-> Checking all 137 links for active status...
Keeping unverified (HTTP 403): https://www.glassdoor.com/job-listing/j?jl=1010301949473
Keeping unverified (HTTP 403): https://www.glassdoor.com/job-listing/j?jl=1010301526719
...
* 137 active postings remain (90 unverified)
[4/5] Scoring against resume...
-> Parsing resume: Alex_Rivera_Resume.pdf
-> Extracted 46 keywords: Westlaw, LexisNexis, e-discovery, Contract Review, Legal Research, Litigation Support, relativity, document management...
-> Scoring 137 roles against keywords...
* Scored all 137 roles (range: 3% — 92%)
[5/5] Generating reports...
-> Building PDF: 4 pages (137 roles)
* PDF saved: results\paralegal_nyc_2026-10-21-050656.pdf
-> Building CSV: 137 rows, 36 columns
* Tracking: ready to fill in Excel, Sheets, or Notion
* CSV saved: results\paralegal_nyc_2026-10-21-050656.csv
================================================================================
Sweep Complete!
Active roles found: 137
Top match: Litigation Paralegal at Hoagland Longo Moran Dunst Doukas (92%)
Reports saved to: results
================================================================================
A sweep takes a few minutes, depending on your internet speed, cache state, and platform count.
Sweep stages
| Stage | What happens | What to watch for |
|---|---|---|
| Profile loading | Reads your YAML config file | Confirms your titles and location match your profile |
| Discovery | Sends search requests to each applicant tracking system platform | Shows how many raw postings each platform returned |
| Validation | Checks every link, then removes dead postings | Keeps unverifiable postings and marks them unverified |
| Scoring | Extracts terms from your resume and scores each role | Shows the term count and score range |
| Generation | Builds the PDF card directory, the CSV export, and any other enabled format | Displays the file paths where the engine saves your reports |
Sweepy deletes a posting when the platform returns a hard 404 or a soft-404 page, and when the connection fails.
Rate limits, bot blocks, transient server errors, and timeouts prove nothing about the posting.
Those cases return 401, 403, 408, 425, 429, 500, 502, 503, or 504, or no response at all.
Sweepy keeps the posting in the run and marks it unverified.
Locate your files
When the sweep finishes, open the results folder in your Sweepy install.
Get-ChildItem results
The directory shows files stamped with today's date and time. The filename records the run time on a 24-hour clock. Windows shows the file time in your local time zone, so the clock values can differ by a few hours:
Directory: C:\Users\Alex\sweepy\results
Mode LastWriteTime Length Name
---- ------------- ------ ----
-a--- 10/21/2026 1:06 AM 19065 paralegal_nyc_2026-10-21-050656.pdf
-a--- 10/21/2026 1:06 AM 215694 paralegal_nyc_2026-10-21-050656.csv
Double-click the PDF file in File Explorer, or run this command in your terminal to open it immediately:
start results\paralegal_nyc_2026-10-21-050656.pdf
Running daily sweeps
Sweepy supports daily sweeps. Run the same command every morning to catch new postings:
sweepy --run
Each run that finds new postings stamps the filename with the date and time, so a new run never overwrites your previous results. When nothing new appears, the run prints No new postings since your last sweep and leaves results untouched. Over time, your results folder builds a complete history of your job search.
On Windows, you can create a scheduled task to run the sweep every morning at 8:00 AM. On macOS/Linux, use a cron job. The command-line reference covers both approaches.
You finished your first automated sweep. Next, open the PDF. Review the match scores and the color badges.