Journalists collect contacts constantly: press releases, PDF report appendices, corporate about pages, court filings, and long email threads with sources. Building a press list or tracking down a spokesperson means pulling addresses out of documents that were never meant to be a database.
This workflow helps reporters and researchers extract every address from pasted documents and pages, filter for official or institutional domains, and keep a tidy contact sheet - with all processing local so sensitive source data stays private.
Journalists & researchers get a cleaner result from “Email Extraction for Journalists” when each step stays separate. Finish the current step and check the output before you open the next tool. The utilities named for this workflow (email extractor, government email extractor, edu email extractor) are linked further down so you can jump to the one that matches the file you have, instead of guessing from the full directory.
Keep the original text or file until the last step looks right. These tools run in your browser, so you can retry a step on the same copy without uploading the list. If the result is shorter than you expected, look back at the input and confirm the addresses, lines, or columns are still there, then rerun only the step that dropped them.
A useful check for journalists & researchers before you export: Keep sensitive source contacts in a separate, access-controlled file - never in a shared press list. After that, make sure the output is the shape the next tool expects, usually one item per line or a single column, and only then import it into a spreadsheet, CRM, or sending platform.
Worked example: copy three rows that look like the real file and run only “Extract from documents and releases”. Compare that small output with what you expected before you process the full list. A common pitfall is accepting the first clean-looking result when a later step still has to remove duplicates, fix separators, or drop blank lines. If the sample is wrong, correct the input and rerun that step. Checklist: keep the original file, confirm the sample output, finish every later step on the sample, then run the same order on the full file and spot-check the first and last rows.