Parseur
Parseur is an AI document parser that turns emails, PDFs, scans, images and spreadsheets into structured data for ops, finance, HR and logistics teams, delivered in real time through a web app or REST API.
What is Parseur?
Parseur is an AI document parsing platform that turns emails, PDFs, scans, images and spreadsheets into structured, production-ready data. It has been in production since 2016 and has processed more than 100 million documents. The publisher, Parseur Pte. Ltd., is a Singapore company founded by Sylvestre Dupont (CEO) and Sylvain Josserand (CTO); it is fully remote, with a team spread across five countries, and has taken no outside investment, describing itself as 100% owned and funded by its founders and profitable since June 2017.
The working unit is the mailbox: one mailbox per document flow, with its own fields and settings. You send documents to it by dedicated email address, drag and drop, REST API or a connector such as Zapier, Make, Power Automate or n8n, then name in plain English the fields you want back. Three engines coexist inside the same mailbox, Vision AI for visual layouts, Text AI for plain text and templates for fixed forms, and the best-suited one is selected automatically. There is no dataset to prepare and no model to train, which is what lets this instruction-based approach absorb layout changes.
Processing runs in four stages: capture, extraction, normalisation and validation against a mailbox-level schema (dates, numbers, addresses, choice lists), then export. Built-in OCR covers 200+ languages, including scans, photos and handwriting, and tables are extracted line by line for invoice items or transactions. Optional Python post-processing is available on Scale tiers. Exports leave in real time through webhooks with retries and delivery logs, live Google Sheets sync, Excel, CSV and JSON downloads, or the 10,000+ destinations reachable via automation platforms.
Governance sits alongside: per-stage logs, alerts on processing, export or quota failure, role-based permissions and audit trails. The announced targets are accounts payable, logistics, HR, real estate and retail. Uptime is claimed at 99.9%+, typically running above 99.98%, with a public status page, and the site displays a 4.90/5 rating over 60 reviews. Site and application are translated into twelve languages.
What it does
- Capture documents through a dedicated email address, drag and drop, the REST API or an integration
- Extract the fields you name in plain English, with no model to train
- Read scans, photos and handwriting with built-in OCR in 200+ languages
- Pull tables apart: invoice line items, transactions, order details
- Normalise and validate dates, numbers, addresses and choice lists
- Export in real time to a CRM, accounting tool, database or spreadsheet
- Download results in Excel, CSV or JSON, and monitor every stage through logs and alerts
When to use Parseur / When not to
A quick filter to help you decide if Parseur is the right fit.
When to use Parseur
- Accounts payable and finance teams keying invoices and purchase orders into their accounting software by hand
- Logistics and supply chain teams processing delivery notes and bills of lading day after day
- HR and recruiting teams that want incoming CVs pushed straight into their applicant tracking system
- Real estate agencies turning inbound lead emails into CRM records in real time
- Non-technical ops teams who need to add or rename an extracted field themselves, and developers who prefer to embed the same parsing through the REST API
When not to use Parseur
- Anyone hoping to chat with a document: Parseur returns named fields and structured records, not a conversation
- Teams that need a mobile app, since the tool ships as a web application and an API only, with nothing on iOS or Android
- Healthcare organisations that require a signed HIPAA business associate agreement, which Parseur does not sign at this stage
- Organisations bound to data residency outside the European Union, or to self-hosting: storage is in the EU and no on-premise deployment is offered
- Very small document flows: Parseur's own pricing page says it plainly, "It looks like you run a tight ship. You may not need Parseur."
How to use Parseur
A typical end-to-end flow, from setup to results.
- Create a free account, with no credit card required; the site claims a first document parsed in under two minutes
- Create a mailbox for one type of document flow, since it carries its own fields and settings
- Send documents in through the mailbox's dedicated email address, by drag and drop in the app, through the REST API, or via Zapier, Make, Power Automate or n8n
- Declare the fields to extract by naming and describing them in plain English, and let the AI identify them
- For fixed layouts, build a template instead; several templates can live in one mailbox and the best match is selected automatically
- Set the normalisation rules: date formats, numbers, addresses and choice lists
- Wire the export: native connector, webhook, live Google Sheets sync, or Excel, CSV and JSON download
- Watch the logs and alerts; documents received past your quota are marked Quota exceeded and can be reprocessed once credits are restored
- Invite team members once you are on a Scale tier or above
Pros & Cons
Pros
- Eleven years of continuous operation and more than 100 million documents processed, run by a company that is profitable and free of investors
- EU data residency, in the Netherlands, even though the publisher is Singaporean
- Complete and public legal documentation: DPA, named subprocessors, GDPR page, named DPO and Article 27 representative
- An explicit contractual clause barring the use of customer personal data to train or fine-tune AI models
- A permanent free plan of 20 pages per month with no credit card, and public pricing readable up to one million pages with a decreasing cost per page
- Grandfathering: existing customers keep their price when the grid changes
- Three complementary extraction engines with no model training, OCR in 200+ languages including handwriting, usable without a developer while still offering a REST API
Cons
- Advanced post-processing and multi-user accounts are reserved for Scale tiers, from 10,000 credits per month, and the Base range caps at a single user
- Unused credits expire at the end of each billing period, and processing stops once the quota is reached until you upgrade, trigger early renewal or buy a top-up
- SOC 2 Type II status is inconsistent: the product site says in progress and on track for 2026, while the help centre says the company is SOC 2 Type 2 compliant
- The ISO 27001, SOC 2 and HIPAA certifications quoted in the help centre are those of Google Cloud Platform, the host, not Parseur's own
- HIPAA is not covered, no business associate agreement being signed to date
- No mobile application: web and API only
- Human support is limited to paid plans and EST business hours, security questionnaires are only completed for Enterprise customers, and displayed prices exclude VAT/GST
Pricing & Plans
Parseur is offered on a freemium basis. A permanent free plan covers 20 pages per month at no cost and without a credit card. The lowest paid entry point is the Micro plan at USD 49.00 per month, or USD 468 per year, for 100 pages per month. Billing is measured in credits, one credit equalling one page: an email or a spreadsheet counts as one credit whatever its length, while a three-page PDF counts as three. Pricing is tiered by volume, with a decreasing cost per page up to one million pages in self-service, beyond which Enterprise quotes apply. All prices are exclusive of VAT/GST, plan changes and cancellation are self-service at any time, and unused credits are lost at the end of each period.
- USD 0 per month - 20 pages per month
- unlimited mailboxes and fields
- AI engines and templates
- 1 user
- 90-day retention
- self-service AI support
- USD 49 per month (USD 468 per year) - 100 pages per month
- USD 89 per month (USD 828 per year) - 300 pages per month
- USD 129 per month (USD 1
- 188 per year) - 1
- 000 pages per month
- USD 269 per month (USD 2
- 388 per year) - 3
- 000 pages per month
- 1 user
- 1-year retention
- no advanced post-processing
- USD 499 per month (USD 4
- 788 per year) - 10
- 000 pages per month - the first Scale tier
- adding advanced post-processing and multiple users
- USD 899 (20k pages)
- 1
- 299 (30k)
- 1
- 599 (40k)
- 1
- 899 (50k)
- 2
- on quote - up to 10 million pages
- unlimited users
- payment by purchase order
- custom contractual terms
- Prices recorded on 13 August 2026 from the publisher's public pricing endpoint
- in USD by default
Data, GDPR & hosting
A consolidated view of how Parseur handles your data.
GDPR overview
GDPR compliance is claimed explicitly, alongside the UK GDPR, the Swiss FADP, California's CCPA/CPRA and Singapore's PDPA. A Data Processing Agreement is published and accepted at signup, and a Data Protection Officer is named and registered: Sylvestre Dupont, at legal@parseur.com. The publisher being Singaporean, the GDPR applies extraterritorially, hence an Article 27 representative, DataRep, covering the EU, the UK and Switzerland with a per-country address table. The lead EU supervisory authority announced is the Irish Data Protection Commission; the local one is Singapore's PDPC. Transfers rely on EU standard contractual clauses, modules 2 and 3, with disputes heard by the Irish courts. Customers get 15 days' notice to object to a new subprocessor, data subject requests are answered within 30 days, breaches are notified within 48 hours of discovery, and Global Privacy Control signals are honoured under the CCPA.
Who owns the data?
You own everything you send. Parseur states that the emails, documents and data you submit remain yours and that you retain all rights to your content. In GDPR terms, the customer is the Data Controller and Parseur the Data Processor for the extraction service. You can delete a document, a template, a mailbox or the entire account from the application; deleted data leaves the servers immediately and is purged from logs and database backups within 45 days at most. Exporting your data in JSON, CSV or XLSX is free throughout the contract and for 30 days after termination. Parseur states that it never sells or shares customer data.
Reuse rights
Extracted data is yours to reuse without asking permission: export it, push it into your CRM, accounting tool or database, or download it in JSON, CSV or XLSX. Parseur processes documents solely to deliver the service and solely on your instructions. Its DPA forbids using personal data to train, fine-tune or otherwise improve any artificial intelligence or machine-learning model serving any party other than the controller, and the site states that customer data is never reused to train its AI models and is never sold. Day-to-day processing is automated, the team declaring that it does not access customer content unless you open an explicit support request. Separately, and as a controller this time, Parseur collects account, billing and newsletter data; payments run through Stripe, so no card details are stored. Five subprocessors are named publicly: ActiveCampaign/Postmark, Google LLC, Mailgun/Sinch, Microsoft and Scaleway.
Data retention & training
Hosting summary
Customer data is stored in the European Union, in the Netherlands, in a data centre described as ISO 27001 compliant, on Google Cloud Platform infrastructure; EU residency is stated on the home page and on the GDPR page. Data is encrypted with AES-256 at rest and TLS 1.2 or above in transit, with SSLv2, SSLv3, TLS 1.0 and 1.1 disabled and certificates issued by Let's Encrypt. Passwords are never stored: PBKDF2-SHA256, a 512-bit salt, one million iterations. Five subprocessors are published with their processing location: Google LLC, Microsoft (backup) and Mailgun/Sinch process in the EU, Scaleway SAS is a French company, and ActiveCampaign/Postmark processes and stores in the United States the emails sent to @in.parseur.com addresses. Transfers outside the EU rely on the Data Privacy Framework and standard contractual clauses. Third-party penetration tests run annually and continuously, a cybersecurity audit can be shared with Enterprise customers, and breaches are notified within 48 hours. The ISO 27001, SOC 2 Type 2 and HIPAA certifications quoted belong to Google Cloud Platform, the host.
Things to keep in mind
Risks and trade-offs to weigh before adopting Parseur.
- The help centre states that Parseur is SOC 2 Type 2 compliant while the product site presents that compliance as in progress and on track for 2026: ask for the current status in writing before relying on it
- The ISO 27001, SOC 2 and HIPAA certifications cited in the help centre are those of Google Cloud Platform, the host, not Parseur's own, and no HIPAA business associate agreement is signed to date
- Unused credits are lost at the end of every billing period, and once the quota is reached processing stops until you top up or move to a higher tier, so a spike in volume can interrupt a live flow
- Only one user is allowed up to and including the Premium tier, which pushes teams towards shared logins unless they upgrade
- Default retention is 90 days on the free plan and one year on the Base range: set it per mailbox if your compliance rules call for something shorter
- The publisher is Singaporean and Singapore law governs the contract, even though the data itself is hosted in the European Union
- One subprocessor, ActiveCampaign/Postmark, processes and stores in the United States the emails sent to @in.parseur.com addresses
Setup & Integrations
Technical difficulty
Low. Day-to-day use requires no technical skills: the interface is point-and-click, there is no model to train and no dataset to prepare, and fields are described in plain English. The FAQ says a first mailbox can be running within minutes, and the features page claims a first document parsed in under two minutes. Ops teams can change fields and instructions themselves, with no IT ticket. An optional technical path exists for developers: REST API, authenticated webhooks and Python post-processing on Scale tiers. Templates take a little more initial setup than the AI engine, in exchange for tighter control.
Deployment
Integrations
Supported languages
Behind Parseur
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Parseur.
Frequently asked questions
Do I need to train a model to use Parseur?
Which documents can Parseur handle?
What exactly is a credit?
Is there a free plan?
What happens when I reach my quota?
Where is my data hosted?
Is my data used to train the AI?
How long are my documents kept?
Do I need a developer?
Is there a mobile application?
Should you pick Parseur?
Parseur is a mature product rather than a recent launch: eleven years in production, more than 100 million documents processed, and a publisher that is profitable and has taken no outside money. That longevity shows in the governance rather than in the marketing. Data lives in the European Union, in the Netherlands, even though the company is Singaporean; the DPA is public, the five subprocessors are named with their processing locations, and a contractual clause bars the use of customer personal data to train AI models.
Two reservations deserve attention. SOC 2 Type II status is stated inconsistently, the product site presenting it as in progress and on track for 2026 while the help centre claims the company is already compliant, so it is worth confirming in writing before relying on it. HIPAA is not covered either: no business associate agreement is signed to date, and the ISO 27001, SOC 2 and HIPAA references quoted in the help centre are those of Google Cloud Platform, the host.
Economically the picture is unusually legible: a permanent free tier of 20 pages per month, a public grid running to one million pages with a decreasing cost per page, self-service upgrades and cancellation, and prices frozen for existing customers when the grid moves. The structural limit is tiering rather than capability, since multiple users and advanced post-processing only arrive at the Scale levels, from 10,000 credits per month, which makes the Base range narrow for a team.
The natural audience is clear: ops, finance, logistics, HR and real estate teams that handle documents in volume and want the extracted fields inside their own systems, without training a model, writing code or filing an IT ticket for every field change.
- Choosing a selection results in a full page refresh.
- Opens in a new window.