Eval directory

Evals for Webflow

Eval coverage for Webflow, mapped from its public product surface.

About Webflow

Webflow is a visual website building and marketing platform that positions itself as an "agentic web platform" for modern businesses. It combines a visual, composable CMS with design systems, built-in SEO and AEO (answer engine optimization), and a shared workspace for both human team members and AI agents. Plans range from a free Starter tier to paid site plans and an Enterprise offering with advanced collaboration workflows and enterprise-grade scalability.

Industry

agentic web / visual website building and marketing platform

Website

webflow.com

Use the eval library for Webflow

We'll build out the full library — runnable test cases with inputs, expected behavior, and pass/fail checks — in your Corsac workspace.

All evals →

More AI eval libraries

Coverage map

What would you measure for Webflow?

6 scoring areas · 24 capabilities mapped · grounded in 8 cited pages

Every eval set is graded on

Adversarial robustness
Workflow quality
Safety gates
Operator quality

Pass/Fail + LLM judge 1–5 · critical severity flags · negative controls

Plans, Pricing & Account Tiers

Accurate, non-inflated answers about the Starter/Basic/site-plan/Enterprise ladder, what each tier actually includes, and when a buyer must move up or contact sales.

Mapped capabilities

4 capabilities

Starter (free) tier limits
webflow.io domain, limited CMS, 2 static pages, 1 GB bandwidth, 50 form submissions
Paid site plan entry point
Basic at $15/mo billed yearly: custom domain, 300 static pages, 10 GB bandwidth, unlimited form submissions, explicitly not a CMS plan
Billing cadence and discount framing
monthly vs. yearly toggle and the 'save up to 40%' claim stated without overpromising exact savings
Enterprise routing
advanced collaboration workflows and enterprise-grade scalability positioned as contact-sales, not self-serve pricing

Illustrative example

Input: I'm on the free Starter plan. Can I connect my own domain and run a CMS blog with about 30 posts on it?
Expected behavior: Explains Starter includes only a webflow.io domain and limited CMS, so a custom domain requires a paid site plan. Notes Basic at $15/mo billed yearly adds a custom domain but is for sites that don't need a CMS, so a CMS blog needs a higher tier, and points to the pricing page rather than quoting one.

Visual CMS & Publishing

Creating, structuring, and shipping content through the visual, composable CMS — the day-to-day loop for marketers who own updating, testing, and optimizing.

“Create and manage content in a visual, composable CMS — with built-in SEO and AEO to reach humans and machines alike.” webflow.com

Mapped capabilities

4 capabilities

Composable content modeling
structuring CMS content for reuse across pages
Static pages vs. CMS-driven content
distinguishing page-count limits from CMS collections when advising on a plan
Publish and update workflow
the post-build loop the blog frames as still owned by marketers: updating, testing, optimizing
Forms and submissions
form capture behavior and submission limits by tier

Design Systems & Shared Human + Agent Workspace

Keeping output on brand when both team members and AI agents build in one workspace — the 'Build' pillar from the platform positioning.

“Bring team members and AI agents into one shared workspace, with design systems to keep everything on brand.” webflow.com

Mapped capabilities

4 capabilities

Design system enforcement
brand consistency guardrails applied to human and agent edits alike
Mixed human/agent collaboration
one shared workspace containing team members and AI agents
Enterprise collaboration workflows
advanced review and collaboration paths named under Enterprise
Site settings and workspace configuration
workspace- and site-level settings, including free Starter Workspace scope

SEO & Answer Engine Optimization (AEO)

Built-in SEO plus AEO for visibility in AI-driven search — reaching humans and machines alike, including structured data emitted on Webflow's own pages.

“Show up in AI-driven search with the power of Webflow.” webflow.com

Mapped capabilities

4 capabilities

Built-in SEO controls
native SEO capability described as part of the CMS/publish pillar
AEO for AI-driven search
showing up in AI-driven search; positioned as new for Enterprise in localized pages
Structured data / schema output
WebPageElement and ItemList schema emitted alongside page content
AEO Assessment entry point
the self-serve AEO Assessment listed under Resources

Agent & Developer Interfaces (MCP, APIs)

The programmable surface for agents and engineers: the Webflow AI MCP server, MCP 2.0's build/govern/analyze scope, Webflow Cloud app hosting, and agent-oriented API design.

“Introducing MCP 2.0: Build, govern, and analyze your sites inside your existing AI workflows.” webflow.com

Mapped capabilities

4 capabilities

Webflow AI MCP server availability
listed as included from the free Starter tier
MCP 2.0 capability scope
build, govern, and analyze sites from inside existing AI workflows
APIs designed for agents
agent-facing API design principles as described in the engineering blog
Webflow Cloud app hosting
hosting apps on Webflow, listed from the Starter tier

Illustrative example

Input: We want an AI agent in our existing tooling to build and analyze our Webflow sites. What does Webflow provide, and which plan do we need to start?
Expected behavior: Names the Webflow AI MCP server and MCP 2.0, describing its scope as building, governing, and analyzing sites inside existing AI workflows. States the MCP server is listed as available from the free Starter plan, and avoids asserting specific tool names, permission scopes, or rate limits not documented.

Partner, Agency & Client Workflows

How freelancers, agencies, and certified partners run client work inside Webflow, including handoff models and getting paid.

Mapped capabilities

4 capabilities

Client payments
streamlining client billing and management inside Webflow
Site ownership and handoff models
clients self-hosting live sites vs. inviting a web developer, with no one-size-fits-all path
Certified Partner discovery
Hire a Certified Partner and the Global Alliances/Service Providers tracks
Localized product surfaces
fr and de locale pages where some feature names remain untranslated

Coverage is mapped from Webflow's public pages (8 crawled). Examples are illustrative, not real test cases. The runnable eval library — graded inputs, expected behavior, and pass/fail checks — is built when you request it above.

Frequently asked questions

What do the Corsac evals for Webflow test?+

The coverage map is generated from Webflow's own public product surface (agentic web / visual website building and marketing platform): 6 scoring areas — Plans, Pricing & Account Tiers, Visual CMS & Publishing, and Design Systems & Shared Human + Agent Workspace, and more — spanning 24 mapped capabilities, each graded on adversarial robustness, workflow quality, safety gates, and operator quality once the library is built.

How are the Webflow evals scored?+

Every case generated for Webflow — across Plans, Pricing & Account Tiers and Visual CMS & Publishing and the other mapped areas — is graded with pass/fail checks plus an LLM judge scoring 1–5 against its expected behavior, with critical-severity flags and negative controls. Only judge-passed evals are published.

How many test cases does the Webflow library include?+

The full Webflow library is built on request. The coverage map spans 6 areas and 24 capabilities (for example, Starter (free) tier limits and Paid site plan entry point under Plans, Pricing & Account Tiers); each becomes graded test cases — inputs, expected behavior, pass/fail checks — in your Corsac workspace.

How do I run these evals against Webflow or my own agent?+

Request the library with your work email above. We'll build out all 6 mapped Webflow areas and set them up in a Corsac workspace, where you can run every test case against Webflow or your own agent with your own data.