RFP & Tender Fit Monitor — Bid/No-Bid Scoring & Alerts

RFP & Tender Fit Monitor — Bid/No-Bid Scoring & Alerts

**Watch official government procurement feeds and get a bid/no-bid fit score for every new tender — scored against *your* firm. The monitor checks TED (EU), UK Find a Tender and SAM.gov (US), finds notices that match your keywords / CPV / NAICS codes, then runs a fit engine that returns a go/no-go call, the fit score, eligibility flags, deadline risk and the required docs per notice. Out comes a fit-ranked shortlist your bid team can actually act on — plus an HTML/Markdown digest and optional webhook alert. TED and UK are fully keyless.**

> Paste your keywords and your capability profile, schedule it, and let new tenders come to you already triaged.

Why this monitor wins

• 🔓 Keyless where it counts — TED and UK Find a Tender run with no credentials at all. SAM.gov needs only a free personal API key (skipped cleanly if you leave it out).
• 🎯 Scored against your firm, not generic — every tender is graded versus *your* capabilities, certifications, eligible regions, contract-value band and small-business status. The output is a decision, not a dump.
• 🧮 Deterministic by default, LLM when you want it — a fast, free rules engine scores fit out of the box; flip on optional Anthropic LLM enrichment for messier notice text.
• 🔁 Deltas, not re-dumps — a state store dedupes across scheduled runs, so you see only new/changed notices and only pay to score each tender once.
• 📊 A ranked artifact, not a spreadsheet chore — a fit-ranked HTML + Markdown digest is written to the Key-Value Store (DIGEST.html / DIGEST.md) for the bid team to read.
• 📬 Push alerts — optional webhook fires a compact JSON run summary to Slack/Make/Zapier/webhook.site when a run finishes.

What it monitors

Input

• sources — feeds to check: TED, UK_FTS, SAM_GOV (default ["TED","UK_FTS"]).
• keywords — terms describing the work you do; filters TED/UK and informs scoring.
• cpvCodes — Common Procurement Vocabulary codes for EU/UK (e.g. 72000000 = IT services).
• naicsCodes — NAICS codes for SAM.gov (e.g. 541512); one query per code, max 3.
• vendorProfile — your firm: capabilities, certifications, eligibleRegions, minContractValue / maxContractValue, smallBusiness. This is the baseline every tender is scored against.
• generateFitScore — run the bid/no-bid fit engine (default on — this is the product).
• useLlm / llmModel — optional Anthropic LLM enrichment (claude-haiku-4-5 default) when ANTHROPIC_API_KEY is set.
• onlyNewSinceLastRun — emit only deltas vs the previous run (default on).
• lookbackDays — fetch window per run, 1–365 (default 7).
• maxResultsPerSource — cap notices fetched per source per run (default 100).
• weeklyDigest — write the fit-ranked HTML/Markdown digest (default on).
• webhookUrl — optional URL to POST a run summary to.
• samApiKey — free SAM.gov public API key (secret); required only for SAM_GOV.
• fetchSamDescriptions — pull SAM.gov descriptions (extra API call; off by default).
• stateStoreName — named Key-Value Store for the cross-run snapshot.

Output (one record per new/changed tender)

{
  "id": "ted:2026-OJS-123456",
  "source": "TED",
  "title": "Managed IT services and cloud migration",
  "buyer": "Ministry of Digital Affairs",
  "country": "PL",
  "url": "https://ted.europa.eu/udl?uri=TED:NOTICE:123456-2026",
  "deadline": "2026-08-15T17:00:00Z",
  "value": { "amount": 2400000, "currency": "EUR" },
  "cpvCodes": ["72000000"],
  "is_new": true,
  "is_changed": false,
  "fit": {
    "fit_score": 82,
    "bid_recommendation": { "decision": "BID", "confidence": "high" },
    "eligibility_flags": ["region_eligible", "value_in_range"],
    "matched_capabilities": ["managed IT", "cloud migration"],
    "deadline_risk": "comfortable",
    "required_docs": ["ISO 27001 certificate", "financial statements", "references"]
  }
}

Pricing

Pay-per-event. $0.03 per tender fit-scored is the main charge — everything else is small and tied to work actually done:

Tenders are scored once. The cross-run state store means an existing notice is never re-scored or re-charged, so a daily schedule costs you only for what is genuinely new.

Worked example: a daily run over the two keyless sources that turns up 12 new tenders, scored deterministically, with the digest on — $0.01 start + $0.02 sources + $0.036 tenders + $0.36 fit scoring + $0.25 digest = $0.676. Turn weeklyDigest off on daily runs and the same run is $0.426. A quiet day with no new notices costs $0.03.

FAQ

Do I need any API keys? No for TED and UK Find a Tender. SAM.gov needs a free personal public API key, set in samApiKey; omit it and SAM is skipped with a clear message.

How is the fit score calculated? A deterministic rules engine compares each notice to your vendorProfile — capabilities, certifications, eligible regions, contract-value band and small-business status — producing a 0–100 fit score and a BID / NO-BID / REVIEW call. Optional LLM enrichment helps on messier text.

Will it re-send the same tenders every run? No. A named state store dedupes across scheduled runs, so with onlyNewSinceLastRun on you receive only new or changed notices.

Where's the digest? In the run's Key-Value Store as DIGEST.html and DIGEST.md, fit-ranked for the bid team.

Best way to run it? Schedule it (e.g. daily) with a generous lookbackDays; the dedupe store keeps results clean and costs tied to genuinely new tenders.