UI & Frontends
ArrowJS
Agent-era UI framework with WASM sandboxes for safe generated interfaces
Category
UI & Frontends
Interfaces where agents present, generate, edit, or share work with people and other systems.
Editorial scope version 1.1 ยท effective 22 October 2026
Category listing
4 UI & Frontends Tools
UI & Frontends
Agentation
Visual UI annotation and structured feedback for AI coding agents
UI & Frontends
Proof
Collaborative document editor with provenance and agent editing APIs
UI & Frontends
Respira for WordPress
WordPress plugin and MCP server that lets AI agents read and edit live sites in their page builder's own format, staging page and post edits on draft copies by default.
Definition
What belongs in UI & Frontends?
This category covers frontend systems built around agent-generated interfaces, agent-user interaction, or shared human-agent work rather than conventional application UI alone.
Includes agent-generated or agent-editable interfaces, shared human-agent workspaces, and frontends that expose structured state, events, review, or handoff.
Scope
Inclusion and exclusion criteria
Include when
- Agents can create or materially edit the interface or its content.
- The interface exposes structured state, events, or collaboration designed for agent participation.
- Human-agent review or handoff is a documented core workflow.
Exclude when
- It is a general component library or frontend framework with no agent-specific capability.
- AI is limited to copy generation or a superficial chat widget.
Selection guide
How to compare these tools
- What can the agent create, change, or observe in the interface?
- How is state shared between the person and the agent?
- Can agent-originated changes be reviewed or attributed?
Typical use cases
- Let agents generate or update interactive user interfaces
- Share interface state between people and agents
- Review and attribute agent-originated interface changes
Category evidence
Sources for this category guidance
- Agent User Interaction Protocol overview
AG-UI documents a bidirectional, event-based protocol connecting agent backends to user-facing applications, including shared state and human-in-the-loop interrupts.
accessed 2026-09-03
Comparable facts
UI & Frontends comparison
Unknown values are shown explicitly rather than inferred from marketing copy. Open a profile to inspect its claim-level sources.
| Tool | Classification | Pricing | Interfaces | Deployment | Verification |
|---|---|---|---|---|---|
| ArrowJS | Agent-enabling | Open Source | JavaScript and TypeScript packages, Vite project scaffolder, agent skill | client-side browser runtime, server-side rendering and hydration, QuickJS and WebAssembly sandbox | documentation reviewed |
| Agentation | Agent-enabling | Source Available | in-browser annotation toolbar, structured Markdown export, MCP server, webhooks | development-time application embed, local MCP server | documentation reviewed |
| Proof | Agent-native | Free | web editor, HTTP agent API, agent skill, open-source SDK | hosted, self-hosted SDK | documentation reviewed |
| Respira for WordPress | Agent-enabling | Paid | MCP server (npm, local stdio), Remote MCP endpoint on each site with OAuth sign-in, WebMCP, WordPress Abilities, REST API, CLI | self-hosted WordPress plugin, local MCP server and CLI, hosted respira.press account for licensing and OAuth authorization, commercial WordPress plugin; MIT-licensed local MCP bridge (npm) and CLI | documentation reviewed |