Scattered files
Important documents split across drives, email inboxes and local folders, with no single point of access and no search that finds them all.
New: electronic signature and shared knowledge
The company archive that answers, with the exact page.
Theka is AI document management for companies with a lot of documents: contracts, datasheets, manuals, certificates, product photos. It reads them all, tables and images included, and answers you citing document and page.
- Data in the EU
- Bring your own storage
- Bring your own AI key
- Permissions on every answer
Wiki · P200 pump
Topic · product
P200 pump
Stainless steel centrifugal pump for food processing plants. Operates at up to 6 bar1 and must be serviced every 12 months2.
- Certified for food contact (MOCA)3
- Replaces the P150 from 20251
PDF
P200 datasheet.pdf
page 4 · “maximum operating pressure 6 bar”
page 4 · “maximum operating pressure 6 bar”
Needs review: the manual's title says “P210”, but the content describes the P200.
What is Theka
Theka is Italian document management software for companies (DAM) with artificial intelligence. It gathers documents in a versioned archive with permissions and approvals, reads them (text, tables, images, scans) and makes them searchable by meaning. On recurring topics it generates wikis with verifiable citations. It works with your storage (S3, SharePoint, MinIO, SFTP) and with the AI model you prefer.
The problem
Documents everywhere, except where you need them.
In many companies documents are scattered across shared folders, email, SharePoint, the ERP and personal computers. Finding the latest version of a contract, knowing which certificate is still valid or retrieving a figure from a hundred-page manual takes hours. And newcomers spend weeks working out where to look and which document is authoritative.
Uncertain versions
Drafts, revisions and signed copies sit side by side: nobody is sure which is the current version, or who approved it.
No trail
Who read, edited, downloaded or shared a document? Without a reliable log, compliance becomes a risk.
The product
Three things, done well.
An orderly archive, knowledge that builds itself, complete control over who sees what.
Document archive
Every document with its versions, its status and its history.
- Versions and lifecycle (draft → signed)
- Multi-step approvals
- Categories, types, custom fields
- Customers, suppliers, products as master data
Knowledge from documents
Theka finds the topics that keep coming up and writes their wiki.
- Facts extracted with document and page
- Contradictions between documents flagged
- Chat that answers citing its sources
- Shared or restricted, by category
Control and security
Every search and every answer respects the permissions of whoever is asking.
- Roles and permissions by category
- Audit log of every action
- Microsoft and Google SSO, MFA
- Built-in electronic signature
How it works
You upload. Theka reads. You ask.
- 01
Upload or connect
Drag your files in, connect SharePoint, S3 or an SFTP folder, or send them via API from your ERP. Even thousands of documents at once, with a time and cost estimate before you start.
PDFOfficeImagesScans - 02
Theka reads everything
Text, tables, charts and photos, with OCR where needed. It fills in the fields, writes the summary, finds the topics and flags when content and title don't match.
OCRTablesImagesClassification - 03
Find and ask
Search in your own words, not by file name. The answer comes with document and page, and only from the documents you are allowed to open.
Search by meaningChatWiki
The difference
The same archive, two ways of living with it.
On one side, the traditional company archive made of folders and email; on the other, document management software that knows what is inside the documents.
Without Theka
- Documents scattered across shared drives, email and local folders
- Several versions in circulation, none of them certainly valid
- Search by file name only
- Knowledge lives in people's heads
- Approvals by email, with no trail
- Signatures collected by hand and signed documents to be filed again
- Answering a question takes hours
With Theka
- A single digital archive, by category, project, customer and product
- Versions with a clear status, from draft to signed
- Search by meaning inside text, tables and images
- Wikis generated from documents, with citations
- Multi-step approvals, all recorded
- Electronic signature in the document, with the signed PDF already filed
- The answer in seconds, with document and page
Search and AI
Not a list of files. The sentence you were looking for.
Theka combines seven signals (words, meaning, customers, projects, topics) and shows you the right passage, with the page.
- Finds results even when you don't use the same words as the document
- Reads the numbers inside tables and the data in charts
- Every result respects permissions: nobody sees what they shouldn't
how many bar can the P200 pump reach?
P200 datasheet
Datasheets · PDF · version 3, current
…maximum operating pressure 6 bar at 20 °C…p. 4
…performance table: flow rate 12 m³/h at 6 bar…p. 7
P200 installation manual
Manuals · PDF
…do not exceed a pressure of 6 bar during testing…p. 12
Related topic: P200 pump
3 facts from 4 documents
What's new
What has arrived recently.
Continuous updates, designed with the people who use Theka every day. The details are in the release notes.
Electronic signature in the document
Send a document for signature to several people, in order, and the signed PDF comes back to the archive on its own.
Shared or restricted knowledge
For each category, choose whether its facts join the knowledge of the whole company or stay with those who can open the documents.
“Needs review”
If a document's content contradicts its title, type or category, Theka flags it: you correct it or dismiss it as a false alarm.
Large archives, no surprises
Thousands of files in batches: an estimate before you start, pause, resume and retry only the ones that failed.
Tables, charts and images
Theka reads the data inside tables and charts and describes photos, so that information can be found and cited too.
API for PIM and ERP
Your product catalogue or ERP reads and writes in Theka with a dedicated key and expiring links to files.
AI under control
You choose the AI. And you know what it costs.
Theka doesn't tie you to a provider: you use your own OpenAI, Anthropic, Gemini or Mistral key and decide where and how much it works.
Model by category
An inexpensive model for invoices, a more capable one for contracts.
Three reading levels
Basic, standard or full: more depth only where it is needed.
Daily caps
Spending limits for knowledge and for images and tables. Once the limit is reached, work resumes overnight.
Estimate before you start
Before bulk processing, see how many documents, how much time and how many tokens.
Security
Your documents stay yours.
Security is in the way Theka is built, not in a separate module.
Separate spaces
Each company has its own data space, separated from the others.
Your storage
Files can live in your S3, SharePoint or MinIO. Export and delete whenever you like.
Permissions everywhere
Search, chat and wiki only show what the user is allowed to open.
Everything recorded
Who uploaded, approved, downloaded, signed. Always verifiable.
Who it's for
For those with lots of documents and little time to search for them.
Manufacturing
Datasheets, certificates, manuals and test reports for hundreds of product codes.
Every product has its own wiki: datasheet, certificate and manual are the cited sources.
Firms and services
Contracts, mandates and case files that pile up over the years and get lost every time someone changes role.
“When does the contract with customer X expire?” The answer comes with the clause and the page.
Multiple sites, multiple departments
The same supplier seen by purchasing, quality and logistics.
One body of knowledge, different permissions: everyone sees what they need to.
Glossary
The language of document management, explained.
A short definition to find your way among DAM, digital archive, semantic search, RAG and electronic signature.
- Document management (DAM)
- The software that organises, versions, searches and protects a company's documents. DAM stands for Digital Asset Management; it is also called a corporate digital archive.
- Digital archive
- The full set of a company's documents in electronic form, with versions, permissions and history. It is the foundation for search, approvals and signature.
- Semantic search
- Search by meaning: it finds the right document even when it doesn't contain the exact words of the question, thanks to embeddings.
- RAG
- Retrieval Augmented Generation: the AI first looks up the relevant passages in your documents, then answers using only those and citing them.
- Company knowledge
- The facts extracted from documents and grouped by topic: products, customers, suppliers, procedures. Every fact carries the document and page it comes from.
- Generated wiki
- A summary page on a topic, written by the AI from the facts and updated when the documents change.
- OCR
- Text recognition in scans and photos, which makes digitised paper documents searchable too.
- Simple electronic signature (SES)
- An electronic signature with legal value for many business documents: you send the document, the signers sign online, the signed PDF comes back to the archive.
- Master data
- Customers, suppliers, products and people with their fields. Documents are linked to master data to build an orderly company memory.
- Audit log
- The record of who viewed, edited, downloaded, shared, approved or signed a document, and when.
Frequently asked questions
The questions we are always asked.
Can the AI make up answers?
Every answer and every sentence in the wiki carries a citation to the document and the page. If there is no source, the sentence is not written.
Who sees what?
Search, chat and wiki always respect permissions: a user cannot find or read documents they would not be allowed to open.
Can I use Theka without the AI?
Yes. Knowledge is a module: without it, Theka remains a complete document archive with versions, approvals, master data and signature.
Where is my data stored?
In the EU. Files can stay in your own storage (S3, SharePoint, MinIO, SFTP) and you can export or delete them at any time.
How much does the AI cost?
You use your own AI provider key. You choose the model by category, set daily caps and see the estimate before any bulk processing.
How does the approval workflow work?
For each document type you define the approval steps and who approves: everyone, at least one person or in order. Every decision (approved, rejected, changes requested) stays in the log.
Does Theka also read scans and photos?
Yes. Scanned pages go through OCR, images are described and data from tables and charts is extracted: everything becomes searchable and citable.
Does it integrate with my ERP?
Yes, through the APIs: your PIM or ERP can upload and read documents, with permissions dedicated to the access key.
Keep exploring
The other parts of Theka.
Document management Versions, approvals, master data.Knowledge and wiki Facts, topics, chat with sources.Search and AI Document reading, models, costs.Electronic signature Simple signature inside the document.Integrations Storage, AI, SSO, search, API.Security Permissions, audit log, EU data.Pricing Three flat-fee plans per organisation.
// contact
Let's see it on your documents.
A guided demo with a sample of your archive.