Integration of a company data API
The SIRET is the key, not the company name typed six times
Sirene, Pappers, the company directory and BODACC feed your B2B journeys. Prefill, check, watch: an enrichment policy, not a copy-paste from the registry.
- Sirene connectors in production
- SIRET as key
- watch scoped
What is a company data API and when should you have it integrated?
A company data API gives your application a third party's legal identity from a SIREN, a SIRET or a name: legal name, head office, activity, administrative status. Sirene is the public register. Pappers enriches the file (officers, beneficial owners, documents). The company directory aggregates open sources. BODACC carries legal notices. You have the integration built when the B2B form must prefill, when signup must refuse a struck-off company, or when insolvency proceedings must surface before the next invoice.
The company data APIs we integrate
INSEE / Sirene
The public register. Prefill by SIRET, administrative status, SIRET matching.
Pappers
The enriched file: officers, beneficial owners, accounts, official documents.
Annuaire des entreprises
The open-data aggregate, when you want to stay on public sources.
BODACC
Legal notices: proceedings, sales, creations, to watch in the file.
What a poorly scoped SIRET costs
Open bases exist. Until the label stops being the key, you have three records, and strike-off arrives via a friend.
The sales rep types the legal name six times
A SIRET prefills the form. Fewer errors, fewer duplicates, a record that is comparable from one end of the funnel to the other.
We discover the strike-off after the order
Administrative status is checked at signup and on a watch. An inactive company no longer gets through the funnel.
We have three records for the same SIRET
The SIRET (or SIREN) becomes the key, not the label. Existing-base matching is done with an exception report.
Insolvency news arrives through a friend
Useful notices land in the file. The sales follow-up stops, the risk goes to finance.
What each company data API actually involves
Annuaire des entreprises
data.gouv.frThe open-data entry (annuaire-entreprises.data.gouv.fr) to stay on aggregated public sources. Useful alongside Sirene when you want a more product-like search, without switching straight to a paid enricher. Scoping fixes what you store, and what you look up on the fly.
- Search on public sources
- Complement to Sirene
- Storage minimised
BODACC
ProceedingsThe bulletin of civil and commercial notices. No invented routes here: the integration is judged on the events to watch (insolvency, sale, creation) and on the link to the SIREN already in the base. A notice only has value if it lands in the right file, not in an export read too late.
- Notices attached to the SIREN
- Proceedings visible in the file
- Scope settled at scoping
INSEE / Sirene
Public registerThe public register, operated by INSEE. Ideal to prefill a form and harden a directory: legal name, head office, activity, administrative status. Throughput is limited; the connector is designed with a cache and a queue, not by querying Sirene on every keystroke. Matching an existing base is tested on a sample, with a report of unresolved rows.
- Prefill by SIRET
- Cache and queue under the throughput
- Watch on cessations
Pappers
Enriched fileA company file in one call: identity, officers, beneficial owners, documents. Useful when Sirene is no longer enough and risk (proceedings, officer, accounts) must live in the product. Credit per call and the contractual perimeter are scoped early: you do not query the whole catalogue on every record.
- Enriched file in one call
- Officers and beneficial owners
- Credits scoped, not wasted
Four uses of a company identity
Prefilled B2B form
One SIRET, the record completes. Sales stops playing registry clerk.
Check at signup
Status, form, dissemination: a filter, not licensed KYC. We reject struck-off entities, we do not invent a licence.
SIRET-tagging the base
Catch-up of existing records, exception queue. The label is no longer the key.
Proceedings watch
BODACC and paid sources as a complement. The alert lands in the file, not in a café.
Sales, credit, data quality, legal: what to store
We do not “stick the Sirene API on”. We set SIRET as the key and what you are allowed to keep.
Sales shortens the tunnel
Fewer fields, fewer errors. Commercial time goes back to the conversation.
Credit sees strike-off in time
Watch is a job, not luck. We scope it with risk, not with marketing.
Data quality has a system of record
One record per SIRET. Label duplicates leave as exceptions.
Legal limits what we store
Enriching is not copying everything. The enrichment policy is a deliverable.
What we have shipped, and why it transfers here
The method a company API forces
SIRET as key
SIRET, establishment, merger / closure queue.
Deliverable: identity ruleEnrichment policy
Sources, kept fields, duration, owner.
Deliverable: enrichment policyPublic throughput
Cache, debounce, quotas, fallback if the source is deaf.
Deliverable: throughput strategyThird-party watch
Frequency, BODACC events, recipients, fallback.
Deliverable: watch jobWhat nobody tells you before you sign
Public throughput does not like typing
Sirene is not queried on every letter. A cache and a queue are the connector, not a refinement. Without them, the form falls over at peak hours.
The SIRET is the key, not the label
Comparing legal names creates duplicates. Matching a history is done with an exception report, never as silent fuzzy matching.
Enriching is not storing everything
Officers, beneficial owners, accounts: you do not copy the registry in-house by default. Scoping fixes what the journey needs, and the retention period.
An open source is not an SLA
An incident on the public register is visible. The product must degrade (manual entry, cache) rather than block every signup.
What we measure on a company data project
Company data API integration: your questions
Three steps. First choose the source: Sirene for legal identity, an enricher such as Pappers when the risk file must be more complete, the directory or BODACC as a complement. Then build a connector with a cache, a queue under the throughput, and the SIRET as the key. Finally test the prefill and, if needed, matching of the history. The sensitive part is not the call that returns a legal name, it is the duplicate and the watch.
Sirene is the public register: identity, head office, activity, administrative status, under limited throughput. Pappers enriches the file (officers, beneficial owners, documents) and is billed by usage. You often start with Sirene for the form, then add Pappers on accounts where risk justifies it. The two combine; it is not a replacement.
A Sirene prefill costs far less than matching a hundred thousand records plus a BODACC watch. Backfill volume and the level of enrichment weigh more than the vendor. We scope the perimeter up front and give a firm estimate, after a test on a sample if a backfill is planned.
Access and quotas depend on the INSEE journey in force at the time of the project. We recheck them at scoping rather than engraving them here. In every case, the connector isolates secrets if there are any, and behaves under the throughput: cache, queue, form degradation. An integration that assumes unlimited access gets cut off.
In batches, keeping the original input and a report of unresolved rows. We do not silently merge two records whose labels “look alike”. Partial cases (closed establishment, moved head office, homonyms) go into a business queue. SIRET matching is a data-quality project, not a matching script launched on a Friday evening.
Is your base already SIRET-tagged?
30 minutes to compare Sirene vs Pappers, what you store, and how to catch up existing records.
Discuss my company data project