English · 简体中文
xdossier
Turn your AI's specs into HTML that beginners, peers, and experts can all read.
3-tier reader mode · Term popovers · Learning checkpoints · Multi-doc dossier · Single-file HTML · Apache 2.0
What you get
-
3-tier reader mode. Switch between Beginner / Intermediate / Expert in the same HTML, without re-rendering.
Shows term popovers, prerequisite cards, learning checkpoints, or analogies as needed, serving both novices and pros.
- Automatic pedagogy toggles. Term popovers, prerequisite cards, learning checkpoints, and analogies appear or hide based on the reader's level.
- Multi-doc dossier. Cluster specs, changes, and reviews into a unified group based on frontmatter and filenames, generating relation graphs and project-level indexes.
- Single-file output. Double-click to open, works offline, no CDN required. Can be emailed, attached to PRs, or committed to the repository.
Live demos
Real specs rendered by xdossier. Toggle 零基础 / 系统化 / 速查 at the top of any single-doc demo, or open the multi-doc dossier to see relation-graph clustering in action.
MVP-0 Dossier
xdossier organizing xdossier: the MVP-0 spec, 8 implementation notes, and 3 reviews — clustered automatically by frontmatter and filename stems, rendered as one navigable HTML with an inline SVG relation graph. Self-documenting.
Open →Pedagogy Layer Spec
The spec for xdossier's pedagogy layer, rendered by xdossier itself. Shows every feature in one page: reader toggle, prerequisite card, 5 learning checkpoints, 4 analogy callouts, 5 glossary popovers, inline SVG diagrams, comparison cards.
Open →Vision Spec
The foundational vision document. Multi-section spec with semantic blocks (risk register, decision grid, structure map), ASCII-to-SVG architecture diagrams, and 18 section covers.
Open →OSS Launch Research
Competitor analysis, npm name availability matrix, viral launch tactics, naming recommendation. Reference-style content with dense tables and structured findings.
Open →Open the same demo in two tabs. In one, switch reader mode to 零基础 (beginner). In the other, 速查 (expert). Same HTML file. Compare what stays and what disappears.
Why this exists
- The wall-of-text problem. You and your AI agent generate 800+ line specs. A beginner skims and bounces; a senior wants the decision summary. Same doc fails both.
- Markdown can't tier itself. Markdown is linear. HTML carries spatial structure — collapsibles, popovers, sidebars, anchored navigation. xdossier puts that spatial structure to a pedagogical purpose.
- Tiered reading is the missing primitive. No tool today renders the same document for three reader skill levels simultaneously. xdossier toggles inline via a single-file HTML.
- Single-file, offline-shareable. Double-click the HTML; works without network. Share by email, attach to PR, paste into wiki — no CDN, no Docker, no SaaS.
Try it
npm install -g xdossier
xdossier render docs/specs/your-spec.md # opens spec.html with the reader toggle
xdossier cover docs/ # builds dossier covers + project index