BOAMP API: French Public Tenders Scraper (Marchés Publics)

Get French public procurement notices from BOAMP (Bulletin officiel des annonces des marchés publics) as clean JSON: calls for tenders, award notices with winners and amounts, buyer contacts, CPV codes, deadlines, estimated values, lots, tender documents links. Filter by keywords, BOAMP descriptors, CPV, department, notice type, buyer, publication date and response deadline. Turn on onlyNew for daily tender alerts: you only pay for notices you have not received yet.

Price: $1.00 per 1,000 notices ($0.001 per notice). No API key, no login, no monthly rental. Errors, empty results and duplicates are free.

• Source: the official BOAMP open data published by the DILA (French government), about 1.7 million notices, updated every day.
• Covers national notices (MAPA, FNS), European notices published in France (JOUE / eForms) and public service concessions (DSP).
• Every notice is parsed from its full text, whatever its format (eForms, FNSimple or MAPA): you get the same fields every time.
• Ready for AI agents: each notice comes with a markdown summary you can send to an LLM or use as a RAG chunk.

How much does it cost?

You pay per notice delivered, nothing else:

What you get	Price
1 notice	$0.001
1,000 notices	$1.00
A daily alert with ~20 new notices, for 30 days	about $0.60 per month
A one-off export of every IT contract notice for one year (~3,000)	about $3.00

Nothing is charged for errors, for notices dropped by your filters, or for runs that find nothing. If you set a maximum cost per run, the actor stops cleanly when it is reached and never delivers free "preview" items.

Input

All fields are optional. Filters are combined with AND, and the values inside one filter with OR.

Field	Description	Example
keywords	Words searched in the notice title (accents and case ignored)	["informatique", "logiciel"]
keywordsMatch	any or all	"any"
keywordsScope	title (recommended) or fullText (whole notice)	"title"
strictKeywordMatch	Drops titles that only share a word root ("informations" for "informatique")	true
excludeKeywords	Drop titles containing these words	["travaux"]
noticeTypes	APPEL_OFFRE (contract notice), ATTRIBUTION (award), RECTIFICATIF, ANNULATION, PRE-INFORMATION, MODIFICATION, EX_ANTE_VOLONTAIRE	["APPEL_OFFRE"]
descriptors	BOAMP descriptor labels or codes ("Informatique" matches both IT descriptors)	["Nettoyage de locaux"]
cpvCodes	Full 8-digit CPV codes	["72000000"]
contractTypes	services, supplies, works	["services"]
departments	French department codes ("01" and "1" both work, "2A", "974")	["31", "81"]
buyerNames	Words in the buyer name	["Toulouse Métropole"]
publishedFrom / publishedTo	YYYY-MM-DD or relative ("7 days")	"2026-10-01"
publishedWithinDays	Shortcut: last N days	7
deadlineFrom / deadlineTo	Response deadline range ("+30 days" works)	"+7 days"
onlyOpen	Only notices whose deadline is in the future	true
onlyNew	Only notices not delivered by a previous run with the same filters	true
maxResults	Maximum notices (newest first)	100
includeNoticeDetails	Parse the full notice (description, CPV, values, lots, winners, contacts)	true
includeRawNotice	Add the complete original notice as rawNotice	false

Example: the 15 latest IT calls for tenders:

{
  "keywords": ["informatique"],
  "noticeTypes": ["APPEL_OFFRE"],
  "maxResults": 15
}

Output

One dataset item per notice. Example (shortened):

{
  "id": "26-100275",
  "url": "https://www.boamp.fr/pages/avis/?q=idweb:26-100275",
  "title": "Prestations de services informatiques",
  "description": "La présente consultation a pour objet la passation de l'accord-cadre visant les prestations de services informatiques...",
  "noticeType": "APPEL_OFFRE",
  "noticeTypeLabel": "Avis de marché",
  "status": "INITIAL",
  "family": "JOUE",
  "familyLabel": "Marchés européens",
  "publicationDate": "2026-10-17",
  "deadline": "2026-11-19T11:00:00+00:00",
  "buyerName": "Saint-Quentin-en-Yvelines - Communauté d'agglomération",
  "buyerAddress": "1 rue Eugène Hénaff",
  "buyerPostalCode": "78192",
  "buyerCity": "Trappes",
  "buyerEmail": "dacp@sqy.fr",
  "buyerPhone": "0139443030",
  "buyerProfileUrl": "http://www.marches-publics.agglo-sqy.fr/",
  "documentsUrl": "https://www.marches-publics.info/mpiaws/index.cfm?fuseaction=dematEnt.login&type=DCE&IDM=1882837",
  "submissionUrl": "https://www.marches-publics.info/mpiaws/index.cfm?fuseaction=demat.termes&IDM=1882837",
  "departments": ["78"],
  "descriptors": ["Informatique (prestations de services)"],
  "descriptorCodes": ["163"],
  "contractTypes": ["SERVICES"],
  "procedureLabel": "Procédure Ouverte",
  "mainCpvCode": "72500000",
  "cpvCodes": ["72500000"],
  "estimatedValue": null,
  "awardedValue": null,
  "currency": null,
  "duration": "48 month",
  "placeOfPerformance": ["FR103"],
  "lots": [{"id": "LOT-0000", "title": null, "description": null, "cpvCode": null, "estimatedValue": null, "currency": null}],
  "lotsCount": 1,
  "winners": [],
  "winnerDetails": [],
  "tendersReceived": null,
  "socialEnvironmentalCriteria": ["environnementaux"],
  "noticeSchema": "eForms",
  "markdown": "## Avis de marché 26-100275: Prestations de services informatiques\n- **Buyer:** Saint-Quentin-en-Yvelines - Communauté d'agglomération\n- **Deadline:** 2026-11-19T11:00:00+00:00\n..."
}

Award notices (ATTRIBUTION) also include winners (names), winnerDetails (name, SIRET, city), awardedValue and tendersReceived when the notice states them. National award notices give the award text in awardDetails.

Use cases

1. Daily tender alert for an IT services company (bid manager)

Every morning at 7:00, a bid manager at an IT services company in Toulouse gets the new IT calls for tenders in Occitanie in Google Sheets, with deadline and DCE link. About 5 to 15 new notices a day, so under $0.50 a month.

Input (scheduled daily):

{
  "descriptors": ["Informatique", "Logiciel"],
  "noticeTypes": ["APPEL_OFFRE"],
  "departments": ["31", "32", "81", "82", "09", "11", "12", "46", "65", "66", "34", "30", "48"],
  "onlyOpen": true,
  "onlyNew": true,
  "maxResults": 200
}

Output (extract):

[{"id": "26-100036", "title": "INFOGERANCE ET MAINTENANCE INFORMATIQUE DES SITES DU GIP FCIP DE TOULOUSE", "buyerName": "GIP FCIP TOULOUSE", "deadline": "2026-11-23T10:00:00+00:00", "departments": ["31", "12", "46", "81"], "documentsUrl": "https://...", "url": "https://www.boamp.fr/pages/avis/?q=idweb:26-100036"}]

Next step: the Google Sheets integration adds a row per notice, and the team qualifies them (go / no go) in a column next to it.

2. Competitor and price watch from award notices (sales director)

A sales director at a cleaning company checks every month which competitors win cleaning contracts in Île-de-France, and at what price. About 300 award notices a month, so $0.30 a month.

{
  "descriptors": ["Nettoyage de locaux"],
  "noticeTypes": ["ATTRIBUTION"],
  "departments": ["75", "77", "78", "91", "92", "93", "94", "95"],
  "publishedWithinDays": 31,
  "maxResults": 1000
}

Output (extract):

[{"id": "26-9xxxx", "title": "Nettoyage des locaux administratifs", "buyerName": "Ville de ...", "winners": ["Société X"], "winnerDetails": [{"name": "Société X", "siret": "123456789000xx", "city": "Paris", "postalCode": "75011"}], "awardedValue": 412000, "currency": "EUR", "tendersReceived": 6}]

Next step: a pivot table by winner gives market shares, contracts won and average contract value; the buyers' names and emails feed the CRM for the next renewal.

3. Lead generation for a public-sector consultancy (business developer)

A consultancy that helps cities with architecture competitions wants every buyer that has launched a project-management (maîtrise d'œuvre) contract worth more than €100k in the last 90 days, with the contact email. About 2,000 notices, so $2 for the whole list.

{
  "keywords": ["maîtrise d'oeuvre"],
  "noticeTypes": ["APPEL_OFFRE"],
  "publishedWithinDays": 90,
  "maxResults": 3000
}

Output (extract):

[{"buyerName": "Mairie de ...", "buyerCity": "Bastia", "buyerEmail": "marches@...", "buyerPhone": "04...", "estimatedValue": 180000, "title": "Marché de maîtrise d'oeuvre portant sur la réhabilitation ...", "url": "https://www.boamp.fr/pages/avis/?q=idweb:26-99420"}]

Next step: filter estimatedValue > 100000 in the spreadsheet, remove duplicates by buyerSiret, then import into the CRM or the emailing tool.

4. AI tender analysis (proposal team)

A proposal team sends the markdown field of every new notice to an LLM, which scores how relevant it is to their offer and writes a 3-line summary for Slack. With onlyNew, a notice is never analysed (or paid for) twice.

Call it from your code

Python:

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("<username>/boamp-tenders").call(run_input={
    "keywords": ["informatique"],
    "noticeTypes": ["APPEL_OFFRE"],
    "onlyOpen": True,
    "maxResults": 50,
})
for tender in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(tender["deadline"], tender["buyerName"], tender["title"], tender["url"])

JavaScript:

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('<username>/boamp-tenders').call({
    descriptors: ['Informatique'],
    departments: ['31'],
    noticeTypes: ['APPEL_OFFRE'],
    maxResults: 50,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
items.forEach((t) => console.log(t.deadline, t.buyerName, t.title));

You can also call it over HTTP with run-sync-get-dataset-items and get the notices in the response.

Automation (Make, n8n, Zapier, Google Sheets)

Daily tender alert in n8n or Make:

1. Create a schedule for this actor in Apify Console (e.g. every day at 6:30), with your filters and "onlyNew": true.
2. In n8n, add the Apify trigger "Actor run succeeded" (or the Make module "Watch Actor Runs").
3. Add Get dataset items with the run's defaultDatasetId.
4. Send each item to Google Sheets (append row: publicationDate, title, buyerName, deadline, estimatedValue, url), Slack or email.

Zapier works the same way with the Apify app ("New finished actor run" trigger). Google Sheets can also be connected directly from the actor's Integrations tab.

Use with AI agents (MCP)

This actor works as a tool in the Apify MCP server, so Claude, Cursor, ChatGPT and other AI agents can search French public tenders on demand ("find open cleaning tenders in Lyon closing next month"). The input is simple JSON, and each result has a markdown field written for LLM context windows and RAG pipelines.

FAQ

Is it legal? Yes. BOAMP data is official open data published by the French government (DILA) under an open licence. The actor uses the public open-data API. It never logs in and never bypasses any protection.

How fresh is the data? BOAMP publishes every day. The open-data API is usually updated the same evening or the next morning.

What is the difference between keywords and descriptors? Keywords search the title. Descriptors are the sector labels given by BOAMP itself (e.g. "Informatique (matériel)", "Espaces verts", "Maîtrise d'oeuvre"). For monitoring, descriptors are usually more complete, and combining both is more precise.

How does onlyNew work? The actor remembers which notices it has delivered for each set of filters, in a named key-value store (boamp-tenders-state) in your Apify account. The next run returns only notices that are not in this memory. It also looks back 7 days, so notices indexed late are not missed. If you change a filter, it counts as a new alert.

Why is a field empty? Notices come in several formats, and many buyers leave optional fields blank (estimated value, email, lots). The actor never makes a value up: a missing value is null.

Can I get DECP data (contracts signed below the publication threshold)? No, this actor covers BOAMP notices only.

Does it include TED (EU) notices? It includes every European notice published by French buyers, because they also appear in BOAMP (family JOUE).

Limitations

• Keyword search uses the BOAMP search engine, which matches word roots. With strictKeywordMatch (on by default), the actor drops titles that only share the root, but a very short keyword ("bus") can still match longer words that start with it ("business").
• CPV filtering needs full 8-digit codes. A prefix such as "72" is not supported by the source.
• Amounts are only those stated in the notice. Many national notices (MAPA) give no estimated value, and some award notices give the amount only as text (awardDetails).
• Older notices (before 2024) use older formats: index fields are always there, but detailed fields can be emptier.
• Departments are those the notice is published for, as BOAMP indexes them. A notice can list several departments.