Docci.ai logo
Document Processing Files · Ocr Doc Parsing

Docci.ai

Docci.ai is an Australian document AI platform that pulls structured data out of PDFs and images, including complex tables and long-form text, then exports it as CSV or JSON through webhooks, an API or Zapier.

Active GDPR compliant Free trial Subscription API available 18+ Verified by Guidaio
Overview

What is Docci.ai?

Docci.ai is a document AI platform published by the Australian company VERMIS AI PTY LTD. Its pitch is narrow and stated plainly on the homepage: stop manual data entry, and pull complex table or long-form text data out of any document. In practice that means sending in PDFs, PNGs or JPGs and getting back structured records you can move somewhere else.

The product is organised around a four-step flow. First you choose a document type, either one of the 100+ preconfigured fieldsets or a custom schema you build yourself in a few seconds. Second you get documents into the system, and here Docci.ai offers more routes than most: drag and drop in the web interface, a connected email inbox that picks up attachments as they arrive, direct API calls, or a bespoke ingestion integration. Third you export in bulk, as CSV or JSON, through webhooks, or into existing systems using custom export workflows. Fourth you search the extracted content from the interface, or query it through the API.

Four capabilities are highlighted around that flow: universal processing across document types, extraction accuracy on form fields and table line items in complex layouts, integration through the API or Zapier, and optional human validation that can be triggered by configured alerts or by low confidence scores. A live demo sits on the homepage — a drag-and-drop zone paired with an extracted-data panel, built around a deliberately awkward invoice.

Four industries are addressed by name: finance and accounting (invoices, expense reports, financial statements), healthcare (patient records, claims, clinical data entry), retail and logistics (supply chain documents, purchase orders, delivery notes), and legal and insurance (contract analysis, claims processing, compliance checks).

What is worth knowing before committing is how little else the vendor publishes. The entire web presence is one marketing page plus three legal documents: no blog, no product pages, no security page, and no API documentation at all, even though API access is a paid feature. Accuracy claims such as industry-leading and unmatched come with no benchmark, and there are no customer names, testimonials or usage figures to weigh them against.

What it does

  • Extract structured data from any document type, including PDF, PNG and JPG files
  • Read complex table line items and form fields out of difficult layouts
  • Apply one of 100+ preconfigured fieldsets, or build a custom schema in seconds
  • Ingest documents through the web interface, a connected email inbox, the API or a bespoke pipeline
  • Export results in bulk as CSV or JSON
  • Trigger webhooks and push extracted data into existing systems via the API or Zapier
  • Route low-confidence extractions to optional human validation
Audience

When to use Docci.ai / When not to

A quick filter to help you decide if Docci.ai is the right fit.

When to use Docci.ai

  • Accounts payable and bookkeeping teams still keying invoices, expense reports and financial statements by hand
  • Healthcare administrators and medical coders processing patient records, claims and clinical data entry
  • Logistics and procurement staff handling purchase orders, delivery notes and supply chain paperwork
  • Insurance and legal support teams working through claims files, contracts and compliance checks
  • Developers and operations engineers who want extracted data pushed into existing systems via API, webhooks or Zapier

When not to use Docci.ai

  • Individuals and very small teams: there is no permanent free plan and the cheapest paid tier costs 149 USD per month
  • Anyone needing a mobile or desktop application, since Docci.ai ships as a web platform only
  • Developers who want to evaluate the API before buying, because no public API documentation is published
  • Buyers who require a published security certification, a DPA or a named hosting jurisdiction before uploading regulated documents
  • Teams expecting instant self-service card checkout: Starter and Business are invoiced manually every month
Get started

How to use Docci.ai

A typical end-to-end flow, from setup to results.

  1. Start the 14-day free trial from the homepage; no credit card is requested
  2. Pick a document type, choosing one of the 100+ preconfigured fieldsets or creating a custom schema
  3. Send documents in by dragging them into the interface, connecting an email inbox, calling the API or wiring a custom ingestion integration
  4. Let the AI read form fields and table line items from the uploaded files
  5. Check the results in the extracted-data panel, using the complex invoice example on the homepage as a reference for what to expect
  6. Switch on optional human validation for configured alerts or low-confidence results
  7. Export the results in bulk as CSV or JSON
  8. Set up webhooks through the API to push data into your existing systems
  9. Connect Zapier if you want to automate the workflow without writing code
  10. Search the extracted content from the interface, or query it through the API
Quick read

Pros & Cons

Pros

  • Pricing is published in full, with figures and monthly page volumes attached to every tier
  • The 14-day trial covers 50 documents and asks for no credit card
  • Four ingestion channels, including a connected email inbox that needs no development work
  • Open output: bulk CSV and JSON export plus webhooks, so extracted data is not locked in
  • Optional human validation catches low-confidence extractions before they reach your systems
  • Customer ownership of uploaded data is stated explicitly in clause 4.1 of the Product Terms
  • No manufactured social proof: no invented reviews, no aggregate rating markup, no fake counters

Cons

  • No permanent free plan, and the entry paid tier costs 149 USD per month
  • API access is sold from the Business plan upwards, yet no public API documentation exists
  • No postal address is published anywhere, and the only general contact is a first-name mailbox
  • No security certification is claimed — no SOC 2, no ISO 27001, no HIPAA — and encryption is never mentioned
  • No DPA is offered, no subprocessor list is published and no hosting country is named
  • The legal documents describe software installed on up to two devices and a comparison viewer, neither of which matches the product being sold
  • Starter and Business are invoiced manually each month instead of being charged automatically
Pricing

Pricing & Plans

Docci.ai does not offer a permanent free plan. The only zero-cost option is a 14-day trial limited to 50 documents, which requires no credit card. The lowest paid entry point is the Starter plan at 149 USD per month for up to 1,000 pages. All prices are quoted in USD and exclude applicable sales taxes; Australian GST is likewise excluded by default under clause 7.1 of the Product Terms.

Free Trial — USD 0 for 14 days, up to 50 documents
  • structured extraction from 100+ preconfigured fieldsets
  • self-serve custom fieldset creation
  • bulk CSV and JSON export
  • no credit card required
Business Plan — USD 599 per month, up to 5,000 pages per month
  • all Starter features
  • basic API access
  • standard support
Enterprise Plan — quote on request, 5,000+ pages per month in custom tiers
  • high volume processing
  • advanced AI capabilities
  • workflow integrations
  • dedicated manager and priority support
  • custom SLAs and user seats
Plan 5
  • Billing note — Starter and Business are settled by manual monthly invoicing rather than automatic card payment
Prices and plans listed above may evolve. Always check the official pricing page before subscribing.
Trust & Privacy

Data, GDPR & hosting

A consolidated view of how Docci.ai handles your data.

GDPR overview

The privacy policy states plainly that Docci.ai collects and keeps information in compliance with both the Australian Privacy Principles of the Privacy Act 1988 and the GDPR, and it lists the usual rights: access, copy, objection, erasure, restriction and opt-out. A Data Protection Officer is designated under Article 37 — the company's General Counsel, reachable at privacy@docci.ai — and lawful bases are set out purpose by purpose. Implementation stops there. VERMIS AI PTY LTD is established in Australia, outside the EU, yet no Article 27 representative is designated. There are no standard contractual clauses, no data processing agreement and no subprocessor list. Transfers rest on a blanket statement that data may be stored anywhere in the world, with consent deemed given by submission. The only supervisory authority named is the Australian Information Commissioner.

Who owns the data?

Under clause 4.1 of the Product Terms you keep all right, title and interest in the data you upload, and you alone answer for its legality, accuracy and quality. Clause 4.2 then grants VERMIS AI PTY LTD a royalty-free, non-exclusive and irrevocable licence to use and reproduce that data in order to deliver the service, improve or enhance it, and for unspecified other development purposes. Clause 4.4(c) casts the customer as data controller and Docci.ai as processor. Feedback is treated separately and less favourably: clause 6.2 lets the vendor use, copy, disclose, licence and exploit it freely, with no royalty and no confidentiality attached.

Reuse rights

Yes. The extracted data belongs to the customer, who can download it in bulk as CSV or JSON, push it through webhooks and feed it into existing systems without asking permission. The restrictions run the other way, against the product rather than the data: the Product Terms forbid copying, modifying, redistributing or reverse engineering the platform (clause 6.3a-b), ban high-risk use in aviation, nuclear, chemical or medical-device settings (6.3d), and — oddly for a tool sold with an API — prohibit using the product through automated means (6.3f). Two further clauses deserve attention: 6.4 lets Docci.ai audit your usage on request, down to the file type and size of each document processed, and 16.3 lets it name you as a customer in its promotional material.

Data retention & training

Retention summary
No retention period is published anywhere. The privacy policy says personal data is kept for as long as reasonably necessary for the purposes it was collected for, plus whatever legal, regulatory, tax, accounting or reporting obligations apply, and may be held longer if a complaint or litigation is in prospect. The criteria listed are the amount, nature and sensitivity of the data, the risk of harm, the purposes, and whether those purposes could be met another way. Erasure can be requested from the Data Protection Officer at privacy@docci.ai, and data identified as coming from a minor is deleted. Nothing is documented about anonymisation, automatic purging, or what becomes of processed documents once an account ends: termination only obliges the customer to destroy or return copies of the software.
GDPR contact

Hosting summary

Docci.ai names no hosting country and no hosting region for customer data. The privacy policy says only that personal data may be used, processed or stored anywhere in the world, including countries outside the European Economic Area, and treats submission of the data as consent to that transfer. The one place where countries do appear — Australia, the United States, or other countries — concerns information gathered by cookies and web beacons, not the documents you upload. Clause 4.4(d) of the Product Terms repeats that personal data may be transferred to or stored outside your own jurisdiction. No standard contractual clauses, data processing agreement or subprocessor list fills the gap. What is anchored is the legal venue rather than the servers: VERMIS AI PTY LTD operates the site from Australia, and both contracts are governed by the law of New South Wales. The domain resolves to a Vercel edge node, which describes the delivery network and says nothing about where documents are stored. Anyone with a data residency requirement will need that commitment in writing before uploading.

Watch-outs

Things to keep in mind

Risks and trade-offs to weigh before adopting Docci.ai.

  • You would be uploading invoices, patient records, contracts and claims to a vendor that publishes no security certification and never once mentions encryption
  • Clause 4.2 of the Product Terms hands the vendor an irrevocable licence over your data, for product improvement and for unspecified other development purposes
  • The privacy policy and the Product Terms disagree on model training: one demands explicit consent, the other never asks for it
  • No hosting country is named, and your documents may be processed anywhere in the world, including outside the European Economic Area
  • Liability is capped at the greater of A$10 and one month of fees, while clause 5.1(d) makes you responsible for every conclusion drawn from the output
  • Automation invites complacency: human validation is optional rather than systematic, and a silently misread invoice line is cheap to create and expensive to unwind
  • Company and product both date from 2025, no address or director is published, and the vendor may terminate without notice, so plan an exit before Docci.ai becomes load-bearing
Setup

Setup & Integrations

Technical difficulty

Low for basic use: start the trial, drag a file into the interface and pick a preconfigured fieldset, or build a custom schema in a few clicks. No technical skill is required. Medium if you connect an email inbox for automatic ingestion. High if you want the API, webhooks or a custom export workflow, which is developer work and is reserved for the Business plan at 599 USD per month. One obstacle applies at every level: no public API documentation is published, so the integration cannot be scoped before contacting sales.

Deployment

Web appAPI

Integrations

Zapier
Company

Behind Docci.ai

Company name
VERMIS AI PTY LTD
Founded
24/05/2025
Country of origin
🇦🇺 Australia
UBO
INFORMATION_NOT_FOUND
UBO country
INFORMATION_NOT_FOUND
Domain registrar country
🇺🇸 United States
Legal contact
Official links

Resources

All the official URLs gathered for verification and reference.

FAQ

Frequently asked questions

What file formats does Docci.ai accept?
The upload area on the homepage lists PDF, PNG and JPG. The vendor states more broadly that its AI works with any document, whatever the complexity or format.
Is there a free plan?
No. The only free option is a 14-day trial capped at 50 documents, and it does not ask for a credit card. Paid plans start at 149 USD per month.
How do I get documents into the platform?
Four ways: upload through the web interface, connect an email inbox so attachments are ingested automatically, call the API, or set up a custom ingestion integration.
In what form do I get the data back?
Bulk export as CSV or JSON, webhook delivery through the API, and custom export workflows into existing systems. Extracted content can also be searched from the interface.
Does Docci.ai have an API, and is it documented?
Basic API access is included from the Business plan at 599 USD per month, and the site repeatedly advertises API and webhook integration. No public API documentation is published, so the integration cannot be reviewed before purchase.
Are my documents used to train the models?
The two legal documents disagree. The privacy policy says uploaded documents are not used to train the models without your explicit consent. Clause 4.2 of the Product Terms grants an irrevocable licence to use and reproduce customer data to improve the product and for other development purposes. Ask the vendor to reconcile the two in writing before uploading sensitive files.
How does billing work?
All prices are in USD and exclude applicable sales taxes. Starter and Business are settled by manual monthly invoicing rather than automatic card payment, and Enterprise pricing is quoted on request.
Who is behind Docci.ai, and which law applies?
VERMIS AI PTY LTD, an Australian company. The site states it is operated and provided from Australia, and both the Terms of Use and the Product Terms are governed by the law of New South Wales.
Is there a minimum age?
Yes. The privacy policy states that the products and services are not designed for and not marketed to people under the age of eighteen.
How do I contact the team?
By email at wilmer@docci.ai for general matters, legal@docci.ai for the terms and privacy@docci.ai for data protection, or by booking a call through the scheduling link on the site. No postal address or telephone number is published.
Conclusion

Should you pick Docci.ai?

Docci.ai does one thing and describes it clearly: it turns documents into structured data and hands that data back in formats you can actually use. The four-step flow is easy to follow, the four ingestion channels are genuinely more generous than the single upload box most competitors offer, and the connected email inbox removes development work from the equation entirely. Pricing is published in full, with page volumes attached to each figure — still not a given in this market — and the 14-day trial asks for no card. The vendor also resists the usual temptations: there are no invented testimonials, no aggregate rating markup and no unverifiable user counters.

The reservations are about what is missing rather than what is claimed. The entire public presence is one page plus three legal documents. API access is a paid feature with no documentation behind it, so a developer cannot scope the integration before signing. No postal address, no named director, no security certification, no mention of encryption, no DPA, no subprocessor list and no hosting country — for a service you feed with invoices, patient records and contracts, that is a long list of blanks. The legal documents are Docci's own but describe another product entirely, one installed on two devices with a comparison viewer, and the privacy policy and Product Terms contradict each other on whether customer data trains the models.

The sensible approach is to trial it on real documents, judge the extraction quality yourself since the accuracy claims carry no benchmark, and get written answers on training, hosting and security before any regulated document leaves your network. As a back-office tool for a team drowning in manual data entry, it is worth the fortnight. As a system of record, it is not there yet.