Skip to main content

​
Overview

Spur AI is a built-in chat assistant available across the entire Spur app. Describe what you want to investigate, organize, or generate in plain language, and Spur AI works through it using your tests, suites, and run data. There are two ways to use it:
  • Chat widget — a persistent widget available on every page. Dock it to the side or float it as a compact panel. The widget picks up context from the page you’re on, and your chat session follows you between pages.
  • Full-page view — click Spur AI in the left sidebar to open the dedicated chat page with quick access to your recent sessions.

​
What you can do

Summarize recent runs

Ask for a summary of your recent runs to see how your tests have been performing without digging through run history.

Triage and heal failing tests

Work through failures conversationally — Spur AI surfaces screenshots and logs so you can understand what went wrong and fix it.

Generate and organize tests

Generate new tests from the material you provide, or ask Spur AI to help organize your existing tests.

Track coverage and test health

Ask about your test coverage and overall test health to spot gaps and flaky areas.

​
Sessions and context

Conversations are saved as sessions, so you can pick up where you left off. Recent sessions are one click away from both the widget and the full-page view, and an active session persists as you navigate — the widget carries your conversation from page to page and uses the current page as context for your questions. You can give Spur AI extra context by attaching material to your message:
  • Spreadsheets — upload a CSV or spreadsheet, for example test input data
  • Documents — upload PDFs such as specs or requirements
  • Jira tickets — reference a ticket to give Spur AI its details
  • PRs or commits — point Spur AI at a code change
  • Loom links — share a recording as context

​
Rich results

Spur AI renders results inline in the conversation — charts, screenshots, log viewers, and test run cards — so you can inspect outcomes without leaving the chat. For multi-step requests, a todo panel shows the tasks Spur AI is working through.

​
Best practices

  • Reference tests by name — the closer your phrasing is to the actual test or suite name, the more accurate the match.
  • Name environments explicitly — include the environment (for example, “on staging”) when your request targets a specific one.
  • Follow up on failures — after asking what failed, ask how to fix it to get debugging context from logs and screenshots.
  • Attach data for context — if your request depends on external material, attach the spreadsheet, document, ticket, or PR rather than describing it.