Live runtime status
GitHub-backed WASM bootstrap
The page is intentionally honest. If GitHub does not actually host the generated JS glue and WASM binary, it shows the precise failure instead of pretending the web surface is published.
Endpoint diagnostics
GitHub-backed URLs only
What is different from Mermaid?
The project is not trying to be a browser-only clone.
Frankenmermaid runs parse → deterministic layout → renderer surfaces from one Rust-first pipeline, instead of a browser-first engine with secondary integrations.
The WASM surface includes sourceSpans() and structured parse results, making editor tooling and click-to-source workflows
possible without scraping rendered output.
Support is tracked with a capability matrix instead of vague blanket claims. Flowcharts are marked implemented; many diagram families are currently partial by design and are labeled that way here.
The same engine targets CLI, SVG, terminal, canvas, and WASM. Mermaid mostly stops at browser rendering; frankenmermaid is being built as a shared rendering core.
Showcase atlas
Big, realistic diagrams first. Everything else is secondary.
This section leads with large live renders from the frankenmermaid browser runtime. It spotlights one serious sample from every major diagram family so the page proves visual output immediately instead of burying it under tooling panels.
Loading spotlight…
pending categoryRendered gallery
80-example render atlas
Browse the whole corpus by category or search term. The renders stay large, the code stays collapsed until you need it, and any card can jump back into the spotlight or the editor.
Live playground
One editor, two engines, six API surfaces
Select a sample, then inspect type detection, parse output, source spans, the capability matrix, SVG output, and the canvas surface. Mermaid is rendered on the right as a baseline comparison.
Syntax lens
Productivity affordances
Theme and style studio
Telemetry JSON will appear after the next committed render.
Determinism evidence JSON will appear here after the checker runs.
Support and evidence
Support truth, matrix coverage, and project reading guidance
This section is the due-diligence rail for the showcase: capability truth, scenario coverage, and reading guidance should stay reviewable without displacing the primary visual story above.
Support reality
Current diagram support from the checked-in capability artifact
Render notes
How to read the current state of the project
They are the only family marked fully implemented in the capability matrix, so they are the clearest way to judge the renderer today.
Sequence, class, state, ER, gantt, and journey are intentionally visible here because the point of the showcase is to reveal current quality, not hide it.
The gallery above is now the primary product surface. The editor below exists so you can inspect parse output, spans, canvas behavior, and Mermaid comparisons after you have already seen the renders at full size.