florida-tenders-monitor-stub

A free, one-off Python script that pulls a small sample of Florida state-agency procurement advertisements from MyFloridaMarketPlace's Vendor Bid System (vendor.myfloridamarketplace.com) - the same public search endpoint the portal's own search page uses - and saves them to a local JSON file. It runs once, does exactly one thing (fetch page 1 of currently OPEN advertisements), and has no memory of previous runs: run it twice and you'll pull the same current snapshot twice, with no idea what changed in between. It is meant as a honest, inspectable starting point for anyone who wants to see the real MFMP data shape before deciding whether they need the scheduled, delta-aware, production version of this scraper.

Setup and run

pip install -r requirements.txt
python main.py

This sends one POST request to https://vendor.myfloridamarketplace.com/mfmp/pub/search/bids for page 1 of status: ["OPEN"] advertisements, prints the first record to the console, and writes up to 20 records to florida_tenders_sample.json in the current directory.

Example output

florida_tenders_sample.json is a JSON array of raw advertisement objects, exactly as the portal's search endpoint returns them (no reformatting). This is a real record pulled by running python main.py against the live endpoint:

{
    "agencyAdNumber": "DOT-ITB-27-9011-CA",
    "type": "Agency Decision",
    "title": "GRASS SEEDS",
    "openDate": "2026-09-08T21:00:00.000+00:00",
    "closeDate": "2026-09-11T21:00:00.000+00:00",
    "favorite": false,
    "status": "OPEN",
    "typeId": "1",
    "version": 1,
    "advertisementId": 16900,
    "uniqueName": "AD-16900",
    "publishDate": "2026-09-08T18:45:42.000+00:00",
    "organization": {
        "organizationId": 30000021,
        "entity": "550000",
        "shortName": "FDOT",
        "vbsAgency": true,
        "version": 0,
        "name": "Florida Department of Transportation (FDOT)"
    },
    "agency": "Florida Department of Transportation (FDOT)"
}

That's the listing endpoint's raw shape: dates as the portal's own raw timestamps (no UTC/Florida-time normalization, no daysUntilClose), agency identity duplicated both flat (agency) and nested under organization, and no description, commodityCodes, docs or responseContact - those only come back from a separate per-advertisement detail request, which this stub does not make.

What this doesn't do

This script deliberately stops at "fetch and save." It does not:

• Schedule itself. There's no cron, no Apify schedule, no daemon - you run python main.py and it does one pull, right now.
• Track deltas between runs. It doesn't remember which advertisementId or version it saw last time, so it can't tell you what's new, amended (version rose) or changed status since your last run - every run is a blind snapshot.
• Retry on failure. A timeout, a dropped connection, or an HTTP 429 from the portal just prints an error and exits. No backoff, no re-attempt.
• Handle dead letters. If a request fails partway through a larger pull, there's no queue of failed records to retry or inspect later - it's just gone until you run the script again.

For scheduled runs, delta/change-tracking, and reliability guarantees, see the production actor: https://apify.com/stefano_seggio/florida-tenders-monitor

License

MIT - see LICENSE.