Skip to main content

Built for documents you can't afford to get wrong

ProcuLink handles the documents your business runs on. We treat that responsibility, and your data, with the seriousness it deserves.

How your documents are protected

  • Encryption in transit and at rest

    Delivery credentials are encrypted at rest with AES-256-GCM authenticated encryption and are never written to application logs. Every outbound channel configured by URL — webhook and HTTP delivery, and S3 ingestion — must use https://, and a plain http:// address is refused when you save it. SFTP is encrypted by construction, and FTPS negotiates explicit TLS. Two connection settings can still weaken that if you switch them on: skipping FTPS certificate validation, and turning SSL off for IMAP or SMTP.

  • Where your data lives

    Your files, the database behind them, and the API all run in EU-region infrastructure — Cloudflare R2, Neon, and Railway. Some processing runs on named US subprocessors under standard contractual clauses: sign-in, AI document extraction and mapping suggestions, payments, and email in both directions — the documents sent to your flow's address, and the data we email out. If you deliver by email, the output itself passes through a US provider. Each one is listed with its location and contract on /subprocessors. Where your data is stored and the route it travels are two different questions: the network path out to your supplier is chosen by our hosting provider and is not pinned to a region by us, so we cannot tell you today which country an outbound delivery leaves from.

  • What we record for every delivery

    Every read, edit and delivery is recorded. Each delivery keeps its channel, how many attempts it took, the last response code and, when it fails, the error. You can open any single document and read its complete history.

  • Validation before delivery

    Built-in checks run on every plan and cannot be switched off: we flag a value that is missing, in the wrong shape or hard to read, and we hold a document we have already seen instead of sending it twice.

  • Access control

    Org-scoped data isolation on every query, scoped API keys you can revoke instantly, and short-lived sessions by default.

  • Responsible AI

    Suggestions never apply themselves: you accept each one, and every value shows how sure we are. Your data is never used to train third-party models. We can also turn AI off for a workspace entirely, so nothing about its documents reaches OpenAI. Documents there cannot be read, because reading uses AI: they are stopped and marked as not read, never sent out. We set that for you on request; it is not a self-serve setting, and it does not change where your files are stored.

Compliance

GDPR
DPA published · SCCs for transfers outside the EEA
SOC 2
Readiness on our roadmap
ISO 27001
On our roadmap
Read the data processing addendum

Subprocessors

CompanyWhat it does for usWhere
RailwayAPI and background-worker hostingEU region
NeonPostgreSQL database hostingEU region
CloudflareR2 object storage (the documents you send us and the output files we generate), DNS, and the relay that passes inbound email to our APIEU-region bucket; the relay runs on Cloudflare's global network
VercelWebsite and app hosting (document files do not pass through it)Global CDN
ClerkAuthentication and session managementUS, EU data residency available
OpenAIAI document extraction and mapping suggestions (API data is not used for model training under OpenAI's API terms)US
StripePayment processing and subscription managementUS, EU establishment
PostmarkEmail in and out: documents emailed to your ProcuLink address, and the files we email to the recipients you setUS
PostHogPseudonymous product analyticsEU (eu.posthog.com)
SentryError monitoring and diagnosticsEU region
Full list with locations, contracts, and change notifications

Need our security package?

We'll share our DPA, security overview, and architecture documentation under NDA.

Found a security problem? Write to security@proculink.eu.