French & EU Public Tenders Monitor (BOAMP + TED)

Track public procurement opportunities across France and the EU from one call, and let AI tell you which ones are worth bidding on. Query BOAMP (the French official procurement journal, Bulletin officiel des annonces des marchés publics) and TED (Tenders Electronic Daily, the EU-wide procurement journal) with keyword, CPV, department and deadline filters, then get one clean, unified, structured record per tender.

Built for bid/tender monitoring, sales & business development to the public sector, market intelligence, and AI agents (works via API and MCP out of the box).

Why this actor

• ✅ Two official sources, one schema. BOAMP (France) and TED (EU) results come back normalized to identical fields, deduplicated, and filtered by deadline.
• ✅ True monitor mode: it remembers what it already reported, so scheduled runs return (and charge for) new tenders only.
• ✅ Slack and webhook alerts. New tenders land in your Slack channel or any endpoint the moment a run finds them.
• ✅ AI relevance scoring, optional and off by default (bring your own Anthropic API key for now; a built-in option is planned). Describe your company once and every tender gets a 0-100 relevance score, a one-line reason, and a bid / consider / skip recommendation.
• ✅ Official open data from DILA (BOAMP) and the EU Publications Office (TED). No scraping, no stale caches.
• ✅ Real filtering on keywords, French department codes, CPV codes, publication window, and "only tenders still open for bids".

The 5-minute tender radar

1. Set your keywords / cpvCodes / departments.
2. Turn on monitorMode and paste a Slack webhook URL.
3. (Optional) Describe your business in companyProfile, add your anthropicApiKey and set alertMinScore: 60.
4. Create an Apify Schedule (daily).

Every morning, your Slack channel gets only the new tenders relevant to your business, scored and sorted. Nothing you've already seen, nothing off-topic.

Use cases

• Never miss a relevant tender. Get every new public contract matching your keywords or CPV codes the day it's published.
• Stop reading irrelevant tenders: the AI score does that reading for you and tells you which ones actually fit.
• Public-sector sales. Build a live pipeline of buyers and deadlines across your segment and region.
• Market intelligence, plain and simple: who's buying what, where, and by when.
• AI agents. Plug into Claude / Cursor via MCP and ask "any open IT tenders in Île-de-France closing this month?"

Input

{
    "keywords": ["logiciel", "cybersécurité", "infogérance"],
    "sources": ["boamp", "ted"],
    "departments": ["75", "69", "13"],
    "cpvCodes": ["72000000"],
    "publishedWithinDays": 7,
    "onlyOpen": true,
    "monitorMode": true,
    "monitorName": "cyber-idf",
    "webhookUrl": "https://hooks.slack.com/services/XXX/YYY/ZZZ",
    "companyProfile": "ESN de 15 personnes à Paris : audit cybersécurité, pentest, développement sécurisé, RGPD. Clients secteur public et santé.",
    "alertMinScore": 60
}

• keywords: matched in the tender object or title (OR logic). Leave it empty and every tender in your other filters comes through.
• sources: boamp, ted, or both.
• departments: French department codes, used as a BOAMP filter.
• cpvCodes: Common Procurement Vocabulary codes, used as a TED filter.
• onlyOpen: drops tenders whose response deadline has already passed.
• monitorMode: only returns tenders not seen in previous runs (tracked per monitorName).
• webhookUrl: a Slack incoming webhook for a digest message, or any URL for a JSON payload.
• companyProfile: enables AI scoring. A few sentences is enough, in French or English.
• alertMinScore: alerts only include tenders scoring at least this value.

Output (one record per tender)

{
    "source": "TED",
    "id": "378060-2026",
    "title": "France, Services de maintenance des systèmes...",
    "buyer": "GRENOBLE ALPES METROPOLE",
    "publicationDate": "2026-06-03",
    "deadline": "2026-07-20",
    "marketType": "Services",
    "cpvCodes": ["72250000"],
    "location": "NUTS FRK24",
    "country": "FR",
    "url": "https://ted.europa.eu/fr/notice/-/detail/378060-2026",
    "relevanceScore": 82,
    "relevanceReason": "Maintenance de systèmes IT pour une métropole : cœur de métier, région couverte.",
    "recommendation": "bid",
    "source_ref": "api.ted.europa.eu (TED / EU Publications Office)",
    "retrievedAt": "2026-07-03T09:12:00.000Z"
}

BOAMP records share the exact same shape (source: "BOAMP", location: "FR dept 75", url to boamp.fr). The relevance* fields appear when companyProfile is set; results are then sorted best-first.

Companion actors (French open-data suite)

• French Company KYC / KYB: enrich any SIREN/SIRET/company name with INSEE SIRENE + BODACC data and qualify the buyers and bidders behind a tender.
• French Real Estate Sales (DVF): official recorded property sale prices straight from the land registry.

Pricing

Pay per event, no subscription:

Event	When
Actor start	Once per run
Tender found	Per tender returned (in monitor mode: new tenders only)
Tender scored	Per tender AI-scored (only when companyProfile is set). Billable event goes live July 18, 2026; scoring is free until then.

FAQ

Where does the data come from? Official open data only: boamp-datadila.opendatasoft.com (BOAMP / DILA) and api.ted.europa.eu (TED / EU Publications Office). Fetched live at run time.

How does monitor mode work? The actor keeps a per-monitorName memory of every tender id it already returned (kept 120 days). Scheduled runs output only what's new, so you're not charged twice for the same tender.

How does AI scoring work? Each tender's key facts (title, buyer, type, CPV, location, deadline) are rated against your companyProfile by a fast Claude model, called with your own Anthropic API key (anthropicApiKey input; get one at console.anthropic.com). You get a 0-100 score, a one-sentence reason, and bid/consider/skip. Token cost on your key is tiny (~$1 per 1,000 tenders), and your key is used for the calls only, never stored.

How fresh is it? Both registries publish continuously; the actor returns what's live at run time. Schedule it daily for a monitoring feed.

Can I filter by amount / procedure type? Keywords, department, CPV, deadline and window are supported today. More structured filters are on the roadmap, so open an issue with what you need.

Does it deduplicate? Yes, within a run (by source + notice id), and across runs in monitor mode.

Does it work with AI agents? Yes, exposed via REST API and MCP, callable as a tool from Claude, Cursor & co.

---

Something missing?

If you need an extra field, another source, or a different output, open an issue on this Actor and describe it. I read every request and small additions usually ship within days. More French company and tender Actors are on my profile.

Keywords: appels d'offres, marchés publics, BOAMP, TED, public tenders France, EU procurement, tender monitoring, tender alerts, veille marchés publics, CPV, appel d'offre API, government contracts, AI tender scoring, bid/no-bid.