AI Elements are patterns for building agentic interfaces. Each one is a composition of Zard UI components that you copy into your project. There is no package to install and no new component to learn. The code is yours.
Patterns
- Chain of Thought A summary that expands into the steps an agent took.
- Chat The thread. Message Scroller keeps your place while the reply streams.
- Checkpoint A point in the conversation you can return to.
- Confirmation A prompt to approve or deny a tool before it runs.
- Context How much of the context window the conversation has used.
- File Tree The files the agent changed, as a folder tree.
- Inline Citation A source in the sentence that opens on hover.
- Message One turn. The person's message, and the agent's reply with its actions.
- Model Selector Two menus. One picks the model the agent runs, the other picks how hard it thinks.
- Plan The steps the agent plans to take, and how far it has got.
- Prompt Input A message field with attachments and a send button.
- Queue The agent's todo list, split into what is left and what is done.
- Reasoning The agent's thinking as it streams.
- Sources The links the agent used for a reply.
- Stack Trace A failed tool call and its stack frames.
- Subagent A subagent run, with its tool calls and result.
- Suggestion A short prompt you can send to the agent in one click.
- Task The steps the agent takes, as it takes them.
- Terminal A command the agent runs, with its output as it streams.
- Test Results A test run by the agent, with its suites and failures.
- Tool A tool call the agent makes, from input to output.
Get started
Add the components
Most patterns are built on Trace. Add it, and any other component a pattern lists on its page.
npx zard-cli add tracepnpm dlx zard-cli add traceyarn dlx zard-cli add tracebunx zard-cli add traceCopy a pattern
Open a pattern above and copy the code of its example into your project. Every example is a standalone component, so it runs as soon as the components it imports are installed. Or ask your coding agent to do it:
Prompt 
Add the Reasoning pattern from https://zardui.com/docs/ai-elements/reasoning.md to my chat.Streaming
The examples that stream, like Reasoning and Task, simulate the agent in the browser with signals and timers. There is no backend and no AI SDK.
To plug in a real model, keep the template and replace the simulated run: write each chunk your stream sends into the same signals. The pattern does not care where the text comes from, only that the signals change.