Frequently asked questions
If it is free, where is the business; what happens to my data and the AI; can I migrate from Orfeo; what AGPL means for a public body.
- Is it really free, or is there a paid edition with the good parts?
- It is free and it is the complete system, artificial intelligence included. There is no cut-down edition. We charge to implement it, migrate data, train, support and host it, which is human work; the software is not.
- If the software is free, how does the project sustain itself?
- Through services. Excribe lives on implementing, migrating, training, supporting and hosting. Free code is not philanthropy: it is what lets an institution audit it, change supplier and avoid being trapped.
- What does the AGPL licence mean for my institution?
- You may use the system for whatever you want, without paying and without asking permission, including modifying it. The obligation arises only if you distribute the modified system to third parties or offer it as a service to other institutions: then you must publish your changes under the same licence. Using it internally creates no obligation to publish anything.
- Will my documents be sent to an artificial intelligence server abroad?
- No, unless you configure it that way for public material. The default runs the models inside your own infrastructure. And material classified as restricted is never sent to an external provider under any configuration: there is no setting to allow it.
- Do I need a graphics card for the artificial intelligence?
- It depends which part you want. Search by meaning and cited precedents use no generative model and run on the processor alone on the minimum server. The conversational assistant does need a card: on the processor it produces two to four words per second, over a minute per answer, and nobody waits for it. With 8 GB of video memory the default model fits and answers drop to a few seconds; with 24 GB a large model fits that drafts regulatory language better. The full table is on the installation page.
- Can I migrate from Orfeo?
- Yes, and we start ahead because Orpyca is its direct heir. But let us be precise: there is no automatic migrator that moves an entire Orfeo installation. What exists is an open import format, and the work is building the translator from your database into it. Records are migrated with their attachments, metadata, reserve classification and history, preserving the original number. Case files, users and passwords do not travel in that package: they are configured separately, and people reset their password on first sign-in. The services page sets it out point by point.
- What if I am coming from a system other than Orfeo?
- It works the same, because the import format is not specific to Orfeo: it is a package with a published schema and integrity verification file by file. What changes is the translator to be written from the source system. In some cases it is cleaner than migrating from a heavily customised Orfeo.
- What do I need in place before installing it?
- Technically: a domain with three names, a TLS certificate, your own reverse proxy — the system does not ship one —, an outbound mail server, which is mandatory, and a configured firewall. Archivally, which is what really decides whether it is useful: a retention schedule approved by the corresponding committee, a classification scheme, a documentary inventory and a defined unit structure. The installation page has the full list with the mandatory items marked.
- Does the system search text inside scanned documents?
- No. There is no optical character recognition: the system stores the image, verifies its integrity and validates the format, but does not extract the text. Search works on what exists as text. If you need to search inside scans, that recognition has to happen before the file is uploaded.
- Is there a mobile app?
- There is no native iOS or Android application, and none is being built. The interface adapts to a phone screen and is used from the browser.
- Does it help me comply with Agreement 001 of 2024?
- Orpyca is designed in accordance with that rule and implements its technical requirements: electronic pagination, an electronic index signed at closure and case file closure rules. That said, an institution’s conformity does not rest on software alone: it also depends on its archival instruments and its procedures. No supplier can certify your compliance, and one offering to do so in writing should raise your suspicion.
- Does the National Archive certify records management systems?
- There is no AGN certification for records management systems. What exists are requirement models published by the AGN and other bodies, against which each institution evaluates the tool it is about to acquire. A supplier claiming to be “AGN certified” is describing something that does not exist.
- Does the electronic signature have full evidentiary weight?
- The system implements advanced signature in XAdES formats with time stamping and certificate status validation. But the certification authority shipped with it is local and not accredited before ONAC: it is good for testing the system, not for attesting. Full evidentiary effect requires connecting an accredited certification authority, and the system explicitly declares that condition on every signature rather than pretending otherwise.
- What happens if one day I want to change supplier?
- We hand over the database and the documents in open formats, at no cost and with no waiting period. Since the software is free, you can keep running the system with another supplier or with your own team. Nothing obliges you to stay.
- How many users does it support?
- The published requirements correspond to institutions of fewer than five thousand users. Above that the system scales by spreading services across more machines; it is worth discussing beforehand to size it properly.
- Can one installation serve several institutions?
- Yes. Each institution has its own database schema and its own object store: data never shares a table. This is the usual way to run it in a regional government with several decentralised agencies.
- What is the difference between Orpyca and Orfeo?
- Orpyca keeps the archival logic of Orfeo and changes everything else: from a PHP monolith to eleven independent services, with per-institution isolation, an unalterable audit trail, advanced signature, preservation conforming to international standards and artificial intelligence. For the counter clerk, the concepts are the same.
- Where is the source code?
- In a public GitLab repository, together with all the technical documentation. You can read it, audit it, copy it and propose changes without asking permission.