Overview
Bugster is an AI-powered end-to-end testing platform that creates, runs, and maintains real-browser E2E and visual tests with self-healing to reduce flaky tests. It surfaces video and log evidence in PRs and runs on every pull request via GitHub App + CLI integration. It generates tests automatically, updates them as flows change, and executes tests in real browsers with rich visual reports, including videos, screenshots, and logs per run. The system uses a two-agent model for Business: an E2E Agent to validate core flows and a Destructive Agent to probe changed areas. Integrations include GitHub App, GitHub Actions workflows, and CLI; supports deployment in Next.js, Vercel, Railway; mentions Claude Code as an agent tooling example. CI covers local, CI, and PR contexts; reporting emphasizes failure evidence and PR comments when failures occur.
Key Features
AI-generated E2E tests
AI agents automatically generate, update, and execute E2E tests for critical flows, with self-healing to reduce flaky tests.
Real browser runs with evidence
Tests run in real browsers and produce video, screenshots, and logs per run.
Two-agent model (Business)
E2E Agent validates core flows and Destructive Agent probes changed areas.
GitHub App + CLI integrations
Integrations with GitHub App, GitHub Actions, and CLI; supports Next.js, Vercel, Railway, and Claude Code.
CI and PR coverage
Run tests locally, in CI, and on every pull request with PR-focused reporting.
Visual reporting and PR feedback
Visual reports, failure evidence, and PR comments when failures occur.
Who Can Use This Tool?
- Development teams:Wants automated E2E and visual testing integrated with GitHub PRs and CI.
Pricing Plans
Starter (free): 0 USD per month; 70 E2E test runs / month; max 5 E2E tests per PR; up to 3 tests in parallel; GitHub App + CLI integration. Good for quick evaluation.
- ✓70 E2E test runs / month
- ✓Max 5 E2E tests per PR
- ✓Up to 3 tests in parallel
- ✓GitHub App + CLI integration
Team (paid): starts at $99 per team per month with usage-based capacity and add-ons; custom E2E run volumes, Destructive Agent runs, increased parallelism, scheduled runs, priority support, and optional SLAs.
- ✓Custom E2E run volumes
- ✓Destructive Agent runs
- ✓Increased parallelism
- ✓Scheduled runs
- ✓Priority support
- ✓Optional SLAs
Coding agents: pay per run model for embedding testing into agent flows; Agent API Plan and partner options available (request access / contact for details).
- ✓Pay-per-run model
- ✓Agent API Plan
- ✓Partner options
Pros & Cons
✓ Pros
- ✓AI-driven E2E testing with self-healing
- ✓Real-browser test runs with video, screenshots, and logs
- ✓GitHub App + CLI integrations
- ✓Visual reports and PR-focused failure evidence
- ✓Two-agent model for Business (E2E Agent + Destructive Agent)
- ✓Broad integrations and CI coverage
✗ Cons
- ✗Public pricing details for Team plan are not published
- ✗No publicly captured Discord invite URL from the site
- ✗Limited external reviews/listings available in quick searches
