Ask your company's documents a question, get a cited answer — without the data leaving your control
EKAP is a self-hosted, open-source AI knowledge assistant that turns scattered internal documents into a natural-language Q&A tool, with every answer citing its source.
Your knowledge is scattered, and search doesn't cut it
PDFs, wikis, spreadsheets, Confluence pages — the answer is in there somewhere, but finding it means asking around or digging through folders. Meanwhile, off-the-shelf AI tools mean sending your internal documents to a third party.
How EKAP works
Connect your sources
PDFs, wikis, spreadsheets, Confluence pages and more feed into EKAP's hybrid retrieval engine, handling everything from quick facts to 300+ page manuals.
Ask a question in plain language
EKAP finds the relevant material and answers like a knowledgeable colleague would.
Get an answer with a citation
Every response links back to the exact document, section and page, so you can verify it.
[Diagram placeholder: architecture/flow diagram — document sources → retrieval engine → access control check → cited answer]
Your data stays under your control
EKAP is 100% self-hosted and open-source, running entirely on your own infrastructure. Every answer is filtered through your existing identity system (LDAP, AD, SSO, MFA) — permission checks happen before retrieval, not after, so people only ever see content they're authorised to see.
External LLM providers are optional and can be gated by document classification when you choose to use them; EKAP doesn't require sending data to a third party by default.
Key features
Hybrid retrieval
Covers quick facts through 300+ page manuals.
Citations on every answer
Exact document, section and page, every time.
Choice of LLMs
Open models, or external providers gated by document classification.
Admin controls
Approval workflow for document changes, nested folder organisation, full audit trail.
Designed for sub-10-second responses, 500+ concurrent users and 99.5% uptime (design targets — pending independent testing).
Deployment and support options
EKAP runs on your own infrastructure. Get in touch to discuss deployment options and support packages for your environment.
Who it's for
IT leads and operations directors at mid-size and larger organisations with internal documents spread across multiple systems, who need staff to get fast answers without giving up control of sensitive data.
Frequently asked questions
Does our data ever leave our infrastructure?
By default, no — EKAP is self-hosted and can run entirely on open LLMs on your own infrastructure. External LLM providers are optional, and can be restricted by document classification if you choose to use them.
How does access control work?
Every answer is checked against your existing identity system (LDAP, AD, SSO, MFA) before retrieval, so users only see what they're already permitted to see.
What document types can EKAP read?
PDFs, wikis, spreadsheets, Confluence pages, and more, via its hybrid retrieval engine.
How fast are responses?
EKAP is designed for sub-10-second responses — this is a design target rather than an independently verified benchmark.
How many people can use it at once?
EKAP is designed to support 500+ concurrent users — again a design target pending independent testing.
Is EKAP certified or compliant with any standard?
We don't claim ISO 27001, GDPR or SOC 2 certification. EKAP is designed to help you keep control of your own data — compliance sign-off is your organisation's responsibility.
Book a demo
A short call to see if EKAP fits how your organisation stores information.