One sovereign workspace
for teams and agents.
Chat, projects, drafts and specialist agents in one place — grounded in your sources, running inside your own perimeter.
A chatbot isn't a deliverable — and it isn't sovereign.
General-purpose tools write a tidy paragraph, then you spend the afternoon turning it into a deck — and compliance won't let you paste the contract in to begin with. Diana is built for the part that actually takes the time, inside your own perimeter.
A secure ChatGPT alternative, on your own terms.
Diana gives your team the everyday usefulness of a chatbot — but as a private, self-hosted AI you run yourself. The language models run on-premise — an on-premise LLM in your own data centre or a sovereign EU cloud — and answer from your own documents, so it's a secure ChatGPT alternative that never sends data outside your perimeter. Compliance teams use it to generate a DPIA, automate their RoPA, and produce cited audit reports — every figure traceable to the source.
A workspace your whole team can open.
Every engagement is a project — files, folders and chats in one place. Share a single document, a whole folder, or the entire project with a teammate, a department or the whole company — or keep it private. Everyone works from the same sources, and Diana works inside the project context.
Share at any level
A document, a folder, or the whole project — shared with one teammate, a department, or the company. Members, status and a due date keep everyone aligned.
Grounded in its files
Drop a project's contracts, filings and notes into folders. Diana reads across all of them and cites the exact file and page.
Diana keeps watch
Suggestions and things that need your attention land in your Inbox — Diana proposes the next move, you review and approve.
Artificial teammates, natural collaboration.
Assign a specialist agent to any document and @mention it in a comment thread — it works inside your project, in context, just like any other member of the team.
It hands you finished documents.
Ask for a memo, a report or a deck and Diana writes it into a real, formatted file — Word, PowerPoint, PDF or Excel. Drafts collect everything you haven't filed into a project yet.
Open any file — Diana's right beside it.
Every document opens with Diana in context. Select a clause to quote it, ask for a plain-language explanation, or quick-edit a paragraph — no copy-pasting into another tab.
Delegate to a specialist agent.
Hand a task to a Researcher, Legal Analyst, Data Analyst or Writer the way you would a teammate. The agent works inside your environment, reads the sources, and reports back — with no external call in normal use.
Just ask — or run a Skill.
Chat in plain language with @ to attach a document, or type / to run a reusable Skill — a legal brief, a board pack, an RFP response — and save the result straight to Drafts.
Built for work that has to be right.
Teams that turn dense source material into finished deliverables — where accuracy, citations and confidentiality aren't optional. See your department →
Legal
NDA and contract review, clause extraction and risk memos — drafted from the data room, with every point cited to the source.
Engineering
Turn a security audit, an RFC or a stack of logs into a clear write-up, a spec or a status report your team can act on.
Marketing
Launch plans, one-pagers and decks — on brand and on message, built from your research and product notes.
Sales
RFP responses, account summaries and pipeline reviews — assembled from your CRM exports and call notes.
From source files to a finished file, in three moves.
Open a project
Drop in the contracts, filings and notes for the engagement. Keep it private, or share it with your team.
Ask Diana
Chat in plain language — or run a Skill. Diana reads across the project’s sources and delegates to agents when it helps.
Ship the document
Get a formatted Word, PowerPoint, PDF or Excel file, cited and finished — saved right where the work lives.
Your week, summarised.
A weekly pulse of every project — what shipped, what needs attention — plus insights on how much each agent is getting done.
And none of it leaves your walls.
Every project, draft and agent runs inside your own perimeter — on-premise, a sovereign EU cloud, or fully air-gapped. No prompts, documents or logs make an external call in normal use — so there's no shadow-AI risk and a clear answer for the regulator.
See it on your
own infrastructure.
Run a pilot inside your perimeter, hold the keys, and let your security team audit every control. Book a briefing.