Verification tools for AI-assisted design.
Every tool here makes one promise with a number in it. We measure the idea before we build the tool, and we measure the tool after. We're building them in the order listed below.
Index
The cards below list each tool's type and surfaces. An agentic loop is AI that works in steps and checks its own output. MCP is a standard way for AI assistants to plug into outside tools. CLI means a command you type in a terminal.
Motif
- type
- agentic loop
- version
- 0.2 · MCP
- surface
- MCP server · CLI · FigJam via Figma MCP
- license
- MIT · data CC-BY-NC
Turns a folder of interview transcripts into findings, then tries to knock them down — and shows you what survived.
Design-system drift
- type
- deterministic check + critic
- status
- not started
Finds where a design file has drifted from the team's design system — its colors, type and spacing, its components — with the receipt — the evidence you can check — for each finding.
Constraint adherence
- type
- deterministic check + critic
- status
- not started
Checks an AI-generated prototype against the rules it was given, and lists what it silently dropped.
Critique
- type
- agentic loop
- status
- not started
A critique tool measured against a known-correct answer before it ships — the Baymard 2023 benchmark, re-run on 2026 models.