One knowledge space for the whole company
The value is in answers, not articles
An employee asks, not searches
"How many vacation days roll over", "who approves a contract", "what to say to a price objection". The answer comes in plain language, with links to articles — you see what backs it up
One topic, one answer
The only thing worse than a missing instruction is two that contradict — an employee follows the first one they find. Gramax spots duplicates and conflicts in the base and shows them to the editor before AI answers from them
You see what changed
An author or the AI agent proposes an edit, and a person reviews it line by line and decides. Accept what's useful, reject what isn't, fix what's debatable. An article never changes behind the owner's back
Knowledge gets checked
You can't ask AI about safety rules or personal-data handling in the middle of a client call — you have to know them. Questions live right in the article: an employee reads the policy and takes the check on the spot
Designed for the AI era
Most systems keep content in a proprietary database. For AI that means narrow REST endpoints, lost structure, and weak context. In Gramax content is Markdown files in Git: the agent reads them in full, with history and a clear structure, the way a developer reads code
| Knowledge base | RAG system | Gramax | |
|---|---|---|---|
| Visual editor | |||
| Collaboration and access rights | |||
| Knowledge storage | |||
| Smart AI search | |||
| Automatic knowledge updates | |||
| Extracting knowledge from meetings, tasks, and chats | |||
| Checking for duplicates and contradictions | |||
| Actions under human control |
Readers don't need licenses
Reader and reviewer roles don't take licenses — everyone can read and approve, licenses are only for authors
Free and Enterprise Versions
Gramax – An Open Source Project
It stays free forever and keeps getting new features
Enterprise Version
For larger teams — centralized access, SSO, roles, and metrics in your own environment