Launch HN: Canary (YC W26) – AI QA that understands your code

Canary is an AI QA agent that analyzes pull requests by understanding codebases, identifying affected user workflows, and automatically generating and executing end-to-end tests. The platform connects to repositories, reads diffs to understand intent, runs tests against preview apps, and reports results directly on PRs with recordings. Beyond PR testing, it supports regression suites and continuous testing via plain English prompts, addressing the gap where modern AI tools accelerate development but lack comprehensive real-world behavior testing before merge.

Key Points

• •Canary reads entire codebases to understand application architecture (routes, controllers, validation logic) enabling context-aware test generation
• •Automatically analyzes PR diffs to identify intent and generate end-to-end tests for all affected user workflows, not just happy paths
• •Executes tests against preview apps and comments on PRs with results, recordings, and flagged behavioral issues before merge
• •Supports multi-modal QA analysis including source code, DOM/ARIA, device emulators, visual verification, screen recordings, and network/console logs
• •Tests can be promoted from PR-specific tests into regression suites and scheduled for continuous execution
• •Enables plain English test creation: describe desired behavior and Canary generates full test suites from codebase understanding
• •Detects second-order effects and edge cases across different user types that standard happy-path testing misses
• •Published QA-Bench v0 benchmark showing 11-26 point coverage advantage over GPT-5.4, Claude Code, and Sonnet on real PRs
• •Requires specialized infrastructure: custom browser fleets, ephemeral environments, on-device farms, user sessions, and data seeding for reliable test execution
• •Solves production incidents by catching regressions (e.g., $1,600 invoice flow drift) before release through comprehensive workflow testing

Workflow Diagram

Concepts