Docs
The manual, kept short.
Install, the workspace, money and keys — honest articles, kept short. The workspace wiki grows one building block at a time.
Every article is a five-minute read — what it does, how to set it up, nothing padded. Filter the lot, or press ⌘K anywhere.
Manuals
how to drive it · 7
Getting started
Install Om-E, connect the Chrome extension, capture your first page and build your first knowledge base.
/docs/getting-started →
Working with chats
The chat manual: attach vs load, comparing sources, citations and verification, editing messages, converting chats to topics, export and import.
/docs/chats →
Working with topics
The topic manual: file imports and image vectoring, transcript ingestion, format topics, drag organising, and sharing topics and KBs as HTML.
/docs/topics →
Working with knowledge bases
The KB manual: one memory per job, meaning-based search, scoped vs universal search, attach vs load, drag organising, exporting a browsable library.
/docs/knowledge-bases →
Files Om-E understands
What happens when you drop a file: supported formats, how spreadsheets become readable — values, formulas, schedules and charts — and the honest edges.
/docs/files →
Settings reference
Every Om-E setting explained: what it does, its allowed range, the default, and when to tune it — Profile, Models, Vectors, Graph, Chat, MCP and Account.
/docs/settings →
Licence & payments
How the Om-E licence works: renting the front end for a term, auto-renewal and how to turn it off, pairing the app to your account, and what happens when a licence ends.
/docs/credits →
How it works
what each part is · 14
Deep research
Multi-source research runs: Om-E searches, reads and verifies across sources, then returns a finished, cited report — saved into your knowledge base.
/features/deep-research →
Web search
Built-in web search — no third-party key. The agent searches, opens the pages, reads them and returns what matters.
/features/web-search →
Citations
Om-E answers cite the topics they came from — the document, the section, the capture date. Hand the answer over and the receipts go with it.
/features/citations →
Verification
Score sources against rubrics you define — security checklists, evaluation criteria, policy rules.
/features/verification →
Comparison
Two proposals, two policies, two vendors — Om-E lays out where they differ and who wins on which criterion, each claim cited.
/features/comparison →
Templates
Templates control the format; your knowledge base supplies the substance. Reports, briefs, memos and SOPs come out finished.
/features/templates →
Agentic
Om-E's agent browses, searches, captures and drafts — one instruction runs a whole errand.
/features/agentic →
Context engineering
A context dial on every turn: attach knowledge for retrieval, load it in full, or keep the chat clean — you control what the model reads and what you pay for.
/features/context-engineering →
Personas
A persona is a standing brief the agent works under — researcher, governance officer, lawyer, coder. The knowledge stays, the eyes change.
/features/personas →
Chats
Chats are where sources meet: attach topics for retrieval or load them whole, compare them, build cited and verified answers.
/features/chats →
Projects
Group the chats that belong to one piece of work. Each keeps its own context — the project builds a shared vector across all of them.
/features/projects →
Knowledge Bases
An organised library of topics, vectored into a private index on your machine — searchable by meaning, nothing uploaded, nothing expiring.
/features/knowledge-bases →
Topics
A topic is one reusable piece of AI knowledge — imported documents with their images vectored too, captured pages, transcripts, notes.
/features/topics →
Multi-session
Kick off multiple chats that run concurrently, each in its own session space — start a research run in one, keep working in another.
/features/multi-session →