Manual data collection from a website (trial task, ongoing work possible)

We need someone to manually look up specific case files on a website (a public court records search tool) and copy the full docket page content into a plain text file, in order from the most recently filed case to the earliest. What the work involves: Searching for specific case file numbers on the site. For each case, copying and pasting the entire docket page as it appears (case number, decedent name, filing date, proceeding type, disposed date, letters status, attorney name, judge, firm, the full party table, and the full list of filed documents) into one continuous plain text file For any case that has a petition document on file, downloading that petition as a Portable Document Format, or PDF, file, named or labeled with its case file number. Solving a simple human verification check that appears periodically during the search process (usually a checkbox, occasionally a harder image based challenge) Working at a steady, natural pace rather than rushing, since this is a government website that asks users to search from a single browser session at a time What we are NOT looking for: No coding, scripting, spreadsheets, or automation of any kind No bots or scraping tools, this must be manual, human lookups No summarizing or reformatting, we want the docket page copied as-is Trial task details: This is a smaller trial batch first, roughly 100 records, paid at a fixed price proportional to a larger ongoing batch of about 3,000 records. We're testing accuracy, pace, and communication before committing to the full batch. What we're evaluating in the trial: Accuracy and completeness of the copied docket text Correctly identifying and downloading petition Portable Document Format files when they exist Comfortable pace without rushing or skipping verification checks Clear communication if a case looks unclear or incomplete Budget: Fixed price for the trial batch, roughly $8 to $10 for 100 records, proportional to a total budget of $150 to $200 for around 3,000 records Note: This work only involves publicly available court filings, no confidential company data.