Import, write, revise, and save a rich document with Markdown and HTML outputs, then export it as Markdown, standalone styled HTML, Word (.docx), or PDF without dropping rich-only formatting.
Where: Flowgen → Add → Words & data → Text Editor
Key ideas
- Editable markdown: Use the inline editor or full-screen split view with a rendered preview for briefs, documents, reports, and reusable context.
- Input is staged: Connected text waits behind Load from input; it never silently replaces local edits when upstream changes.
- Save or export: Save archives/updates a bound generic Brain record, Export PDF prints the rendered document, and md_out sends the current markdown downstream.
- Text Editor card: MarkdownDocNode uses the authored shared-card path. Edit and preview markdown/rich content, stage incoming text, and save to a Brain record.
- Properties sidebar: Card/modal owned; common properties shell
- Opened surfaces: Full-screen rich/markdown editor; AI co-writer; MD/HTML/PDF export
- Rich export path: Word, HTML, and PDF use the rich HTML document when available, so images, tables, colors, alignment, and rich-only content are not silently replaced by plain Markdown.
- Two current outputs: md_out emits portable Markdown; html_out emits the current styled HTML document. Incoming text_in/html_in remains staged until Load from input so edits are not silently replaced.
- Literal Find and Replace: Command/Ctrl-F opens a source-level finder with next/previous match, replacement text, Replace, and Replace all. Matches apply to the current Markdown/HTML source while the formatted preview remains synchronized.
Steps
- Add Text Editor and write directly, or connect upstream text.
- If input is connected, inspect it and choose Load from input only when you want it to replace the current draft.
- Edit in inline/full-screen mode and review the rendered preview.
- Optionally bind/save to Brain or export PDF, then connect md_out to the next node.
- Open the Text Editor, press Command/Ctrl-F or click Find and replace, enter the exact text, step through matches, then replace one or all.
Text Editor controls and presets
- Add path: Flowgen → Add → Inputs → Text Editor.
- Card: MarkdownDocNode.
- Properties: Card/modal owned; common properties shell.
- Controls: Inline/full editor; Rich/Markdown; Load from input; AI instruction/action/model; save/import/export.
- Opened surfaces: Full-screen rich/markdown editor; AI co-writer; MD/HTML/PDF export.
Fullscreen download controls
- Markdown (.md) for portable source.
- Styled standalone HTML with document head, charset, page measure, typography, dark mode, and RTL rules.
- Word (.docx) with heading styles, lists, tables, links, alignment, and embedded/converted images.
- PDF rendered from the same rich standalone document styles.
HTML flow handoff
The HTML socket now propagates the actual current editor content. Connect html_out to HTML Export/raw consumers when a styled finished document is required.
Search behavior
- Literal text matching rather than a regular-expression language.
- Current match is tracked through a replacement so navigation continues after the new text.
- The feature works in the fullscreen editor source surface and updates the node document output.
Rich-document input and editing fixes
PDF and Markdown briefs reach the model through the attachment-preparation path. The text editor defaults body copy to Arial, Find no longer steals the editing caret, and repeated-update failures are corrected. Export still uses the rich document where available rather than discarding images, tables, or formatting.
Limitations and important notes
- Markdown remains the md_out graph output and cannot represent every rich-only formatting detail.
- Word export can warn when a remote or unsupported image cannot be embedded even after conversion.
Troubleshooting
Upstream text changed but my document did not
That is the safety behavior. Choose Load from input after reviewing the staged value; it will replace the current draft only with explicit confirmation.