docs: cover Canvas, Workflow, Gallery, and Model Manager features

[Pfannkuchensack]

Summary

• New: Canvas Workflow Integration (run workflow on raster layer)
• New: Canvas Projects (.invk save/load)
• New: Custom Node Manager (install/manage community node packs)
• New: Layer tips (canvas drag & drop targets + lock transparency)
• gallery: virtual boards (by date)
• editor-interface: workflow library tags (dynamic per-user tags)
• adding-nodes: point manual install at Custom Node Manager
• models: editable source URL, bulk delete/reidentify, missing-files filter, settings export/import
• fp8-storage: hardware support tiers (Ampere / Ada+Hopper / pre-Ampere)

Related Issues / Discussions

Documentation follow-up for the following previously merged feature PRs that shipped without user-facing docs:

• Feature(UI): Canvas Workflow Integration - Run Workflow on Raster Layer #8665 — Canvas Workflow Integration
• feat(ui): add canvas project save/load (.invk format) #8917 — Canvas Project save/load (.invk)
• feat: Add virtual boards that dynamically group images by date #8971 — Virtual Boards
• feat: add Custom Node Manager #9047 — Custom Node Manager
• Feature/user workflow tags  #8698 / Feature: Add Tag System for user made Workflows #8673 — Workflow tags
• feat(MM):model settings export import #8872 / Feat:(mm) model settings export metadata thumbnail #9188 — Model settings export/import (with metadata + thumbnail)
• feat(model_manager): add missing models filter to Model Manager #8801 — Missing Files filter in Model Manager
• feat(UI): add Inpaint Mask as drag & drop target on canvas #8942 — Inpaint Mask drag & drop
• feat(canvas): add per-layer lock transparency option #8943 — Per-layer lock transparency
• feat: add bulk reidentify action for models (#8951) #8952 — Bulk reidentify models
• feat(mm): add editable source_url field to model config #8997 — Editable source_url field
• Feature(UI): bulk remove models loras #8659 — Bulk delete models / LoRAs
• feat: add per-model FP8 layerwise casting for VRAM reduction #8945 — FP8 storage (expanded with hardware tiers)

QA Instructions

1. cd docs && pnpm install && pnpm dev and confirm the site builds without errors.
2. Verify the new pages render and appear in the sidebar:
   • Features → Canvas: Canvas Projects (.invk files), Run Workflow on Canvas, Layer Tips
   • Features → Workflows: Custom Node Manager
3. Verify the modified pages render correctly:
   • Features → Gallery Panel: new Virtual Boards section
   • Features → Workflows → Editor Interface: expanded Workflow Library card with tag filtering
   • Features → Workflows → Adding Nodes: Installing Community Nodes now points at the Custom Node Manager
   • Concepts → Models: new sections Editing model metadata, Bulk actions in the Model Manager, Finding orphaned models, Exporting and Importing Model Settings
   • Configuration → FP8 Storage: new Hardware support tiers section between Requirements and Enabling FP8 Storage
4. Cross-check internal links — Adding Nodes → Custom Node Manager, Custom Node Manager → Creating Nodes guide, Models Finding orphaned models → Bulk actions.

Merge Plan

Docs-only change. No code, no schema, no migration. Safe to merge any time.

Checklist

• The PR has a short but descriptive title, suitable for a changelog
• Tests added / updated (if applicable)
• ❗Changes to a redux slice have a corresponding migration
• Documentation added / updated (if applicable)
• Updated What's New copy (if doing a release after this PR)

[lstein]

• The link to Creating a Node Pack is broken in features/workflows/custom-node-manager/
• You should mention JPPhoto's node connectors in features/workflows/adding-nodes.
• Broken links (x2) to Community Nodes in features/workflows/adding-nodes.
• In concepts/models you do not mention that you can paste a HF RepoID into the URL or Local Path field, and it will be correctly detected.
• In concepts/models do we still need an info box around the Autoimport functionality, which was discontinued way back in v4.0.0?
• In configuration/fp8-storage the important message that FP8 doesn't have an effect on quantized checkpoints comes towards the bottom of the page. I suggest moving the warning earlier, and say explicitly that it is for full precision models.

[lstein]

LGTM.