Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/pewdiepie-archdaemon/odysseus/llms.txt

Use this file to discover all available pages before exploring further.

Odysseus Documents is a multi-tab editor built around a simple philosophy: you write the text, the AI is there to assist — not to replace your voice. You can draft, edit, and organize documents in Markdown, HTML, and CSV, with syntax highlighting and full version history. The AI can make targeted inline edits, propose suggestions as accept/reject bubbles, and help structure content — but only when you ask it to.

Supported formats

FormatFeatures
MarkdownSyntax highlighting, live preview, AI edits and suggestions
HTMLSyntax highlighting, inline preview, AI edits
CSVSyntax highlighting, AI edits
TextPlain text editing, AI edits
EmailCompose window format for drafting email replies

Multi-tab workflow

Open multiple documents simultaneously — each appears as a tab in the editor panel. Switch between tabs without losing your place. Documents are auto-saved and persist between sessions. The agent always acts on the active (currently visible) document. If you have several tabs open and ask “fix the typo in the second paragraph”, the agent edits the tab you’re looking at.

AI features

Inline edits

Ask the agent to edit the open document in chat: “Remove the second section”, “Rewrite the introduction to be more concise”, “Fix grammar throughout.” The agent uses edit_document with precise find-and-replace operations and shows a before/after diff in the tool bubble.

Suggestions

Ask for suggestions or a review — “How can I improve this?”, “Give me feedback” — and the agent responds with inline accept/reject bubbles directly on the document, rather than a prose list in chat. Each bubble shows the proposed change and the reason. Accept or reject each suggestion individually.

AI document creation

When the agent creates a new document (code, report, essay, generated content), it streams the content into the editor in real time so you see it appear line by line, just like a chat response.

File uploads and extraction

Beyond editing text documents, Odysseus can display and extract content from uploaded files:
When the optional PyMuPDF package is installed, uploaded PDF files render in a side panel with per-page display. The agent can also fill PDF form fields directly from the editor.
Note: PyMuPDF is licensed under AGPL-3.0. It is not installed by default. Build with --build-arg INSTALL_OPTIONAL=true in Docker, or install manually with pip install -r requirements-optional.txt.
When the optional markitdown package is installed, uploading .docx, .xlsx, .pptx, .xls, or .epub files extracts the content as Markdown, which then opens as an editable document.

Optional dependencies

The core editor works without any optional packages. The following extras unlock additional capabilities:
PackageFeature unlockedLicense
faster-whisperLocal speech-to-text (microphone → text)MIT
duckduckgo-searchDuckDuckGo as a search provider optionMIT
PyMuPDFPDF page rendering and form fillingAGPL-3.0
markitdownOffice/EPUB document text extractionMIT
To install all optional dependencies in one step:
pip install -r requirements-optional.txt
In Docker, build the image with the optional flag:
docker compose build --build-arg INSTALL_OPTIONAL=true

Document library

All documents are accessible from the Library panel (the book icon in the sidebar). The library shows documents sorted by most-recently updated, with language facets for filtering. Documents can be opened, downloaded, or deleted from the library view. The agent can also list and open library documents with manage_documents — useful for “show me my documents” or “open the Sweden notes doc.”

Build docs developers (and LLMs) love