Design for every agent.
Serve everyone.

A framework for building interfaces that work across the full AI agent capability spectrum —from HTTP retrievers to vision agents to protocol-native systems.

$curl -s https://yoursite.com | grep "<main"
Layer 1: Content accessible to all agents
<div id="root"></div> — invisible to 40%+ of agents

Agent Capability Spectrum

Six levels of AI agents interact with your interface — from HTTP retrievers to protocol-native systems. BiModal Design ensures every level is served.

Learn more

Defense in Depth

Five architectural layers — from server-rendered content through semantic structure, structured data, APIs, to agent protocols — ensure graceful degradation across the entire spectrum.

Learn more

Standards-Based Approach

Built on schema.org, WAI-ARIA, OpenAPI, and emerging protocols like MCP — no custom attributes needed. Use the standards the ecosystem already understands.

Learn more

Performance by Agent Level

Data from WebArena, VisualWebArena, ST-WebAgentBench, and agent protocol benchmarks (2024-2026)

Agent LevelConventional UIBiModal DesignImprovement
HTTP Retrievers (L0-1)12-25%60-75%+40-75%
Browser Automation (L2)35-50%75-88%+30-50%
Vision Agents (L3)40-55%70-85%+25-35%
API / Protocol (L4-5)N/A90-98%New channel
Human Success72-89%72-89%Maintained

Sources: WebArena (Zhou et al., 2024), VisualWebArena (Koh et al., 2024), ST-WebAgentBench (He et al., 2025), Microsoft Build 2025