Honest comparison

Wavemaker Platform vs hosted ComfyUI clouds

Managed Comfy clouds commoditize “run my graph without a GPU.” Differentiation here is operate-in-chat, priced preflight, and registry-honest Trust — not hosting alone.

Product shape

Job Typical hosted Comfy cloud Wavemaker Platform
Run Comfy graphs without local GPU Yes Yes — Modal-backed runtime packs
Operate without opening the graph Usually no — editor-first Yes — guided chat inputs + cost
Itemized GPU + partner API cost Often single GPU meter / opaque Dual-meter preflight before confirm
Copy-ready run API from the UI Varies (deploy-as-API add-ons) Copy API curl for POST /runs + MCP tools
Commercial-use / privacy badges Marketing copy varies Trust Center = public registry fields only

What we claim (receipt-backed)

  • Copy API + dual-meter UI paths are present and unit-covered. — 2026-07-30 Floyo packaging audit PASS
  • OpenAPI platform surface documents runs, files, wallet, and MCP. — docs/platform/guides/api-mcp.md + /developers#platform
  • Node library lists allowlisted packs; rejected packs stay visible (honest empty/reject). — /platform/nodes + packaging audit

What we do not claim yet

Marketing honesty (roadmap D-08). Quality “beats” language unlocks only after golden corpus banking + family bakeoff verdicts.

  • We do not claim broader custom-node coverage than the curated pack seed — Official Comfy Cloud and peers may allowlist differently.
  • We do not claim lower price or faster cold-start than every host; quote the dual meter per run.
  • We do not claim “beats legacy” production quality until corpus + bakeoffs unlock that language (D-08).