AI Interfaces
Tool Retry Policy
A transparent retry sequence showing failures, backoff timing, and stop conditions.
Installation
Copy and paste the code into your project.
Accessibility notes
An ordered list preserves attempt sequence and every failure and delay is readable text.
Retry policy
Webhook delivery
- 1+0s
Attempt 1
Timeout
- 2+1s
Attempt 2
Rate limited
- 3+4s
Attempt 3
Scheduled
Exponential backoff · stop after 4 attempts
Preview accent
const attempts = [{ label: "Attempt 1", result: "Timeout", delay: "0s" }, { label: "Attempt 2", result: "Rate limited", delay: "1s" }, { label: "Attempt 3", result: "Scheduled", delay: "4s" }];
export function ToolRetryPolicy() { return <ol>{attempts.map((attempt) => <li key={attempt.label}>{attempt.label} · {attempt.result} · {attempt.delay}</li>)}</ol>; }<ToolRetryPolicy />Related components
MCP server
Product knowledge
- Transport
- Streamable HTTP
- Round trip
- 84ms
- Search docs
- Read resources
- Run tools
AI Interfaces
MCP Connection Card
A connection summary for MCP servers with transport health, latency, and available capabilities.
Resilience route
Response fallback chain
- 1Ready
Primary · v3
Default
- 2Standby
Fallback · v2-mini
Timeout > 2s
- 3Last resort
Cached response
Providers unavailable
Requests preserve the same output schema across every route.
AI Interfaces
Model Fallback Route
A resilience chain showing primary, fallback, and cached response routes with activation conditions.
Agent activity
3 tool calls
- 1128ms
search_docs
design system tokens
- 242ms
read_file
components/card.tsx
- …running
write_patch
3 changes proposed
AI Interfaces
Tool Call Inspector
A live execution trace for agent tools, file operations, and long-running tasks.