Your QA Stack Costs $212,484 a Year.
NexusQA Costs $3,588.

Triage credits are consumed when NexusQA's AI classifies a bug report — determining severity, priority, category, and suggested assignee. This typically costs 1 credit per triage operation. Remediation credits are consumed when the 7-stage autonomous pipeline attempts to fix a bug — generating a plan, running it through the Opus 4.6 review gate, executing the fix, and verifying it. A full remediation cycle costs approximately $0.50-$1.50 in LLM API tokens, which maps to 1 remediation credit. Starter includes 500 triage credits (no remediation). Professional includes 500 triage + 50 remediation credits per month.

Stage 1: Detection — bugs come from automated tests, human testers, or infrastructure incidents. Stage 2: AI Triage — classifies and assigns. Stage 3: Plan Generation — analyzes root cause using GraphRAG historical context and proposes a fix. Stage 4: Opus 4.6 Plan Review Gate — validates the plan against six criteria including "has this been tried before?", "is this the root cause or a symptom?", and "what is the blast radius?" Stage 5: Execution — creates a branch, commits the fix, opens a PR. Stage 6: Verification — runs regression suite, visual diff, and performance check. Stage 7: Human QA Sign-Off — a tester reviews the PR and approves or rejects. The entire pipeline is transparent — every decision is logged and queryable.

Yes. Each paradigm is a distinct testing methodology with its own tools: (1) Regression — Playwright across 3 browser engines with GraphRAG-predicted execution order. (2) Integration — Pact consumer-driven contracts + Zod schema validation. (3) Smoke — critical-path validation in under 5 minutes. (4) Visual — pixelmatch screenshot comparison with vision-capable LLM analysis. (5) Performance — Lighthouse CI Core Web Vitals. (6) Accessibility — axe-core WCAG 2.1 AA compliance. (7) Security — OWASP ZAP DAST scanning. (8) API — health check monitoring + Zod response validation. (9) Exploratory — human-driven testing with rrweb session recording. (10) Load — k6 and Artillery concurrent user simulation. Each paradigm generates structured data that feeds into the quality scorecard.

When a new deployment arrives with a list of changed files, NexusQA's regression analysis skill traverses the knowledge graph from those files. It finds historically correlated regression patterns — "last time file X changed, tests Y and Z failed." It computes a ranked risk score per test case based on graph path strength and temporal recency, then reorders test execution so the highest-risk tests run first. After each run, prediction accuracy is stored as a graph entity. Over time, the model adapts based on per-model prediction accuracy. This means your most likely failures are caught in the first minutes of a test run, not the last.

NexusQA is designed to augment your QA team, not eliminate humans. The autonomous remediation pipeline handles routine bug fixes — the type of work that consumes 60-70% of QA contractor time. The human tester portal gives your remaining testers a more powerful environment with 8 annotation tools, auto-metadata capture (URL, viewport, browser, console errors), and optional rrweb session recording. For some teams, this means reducing contractor headcount. For others, it means redirecting contractors to exploratory testing where humans excel. We do not make claims about headcount reduction because it depends entirely on your bug volume, codebase complexity, and team structure.

Yes. The price includes all features listed for your tier. No per-seat upcharges for specific paradigms. No "contact us for pricing" on individual features. If you need more triage or remediation credits than your tier includes, you can upgrade tiers or purchase add-on packs at published rates. The one external cost is your own AI provider — NexusQA routes through the Nexus Orchestrator which uses your organization's configured AI provider (Gemini, Claude, OpenRouter). The LLM token cost for AI operations comes from your AI provider subscription, not from NexusQA.

You own your data. On cancellation, you have 30 days to export all test suites, test cases, tickets, screenshots, recordings, and metrics via the API. After 30 days, all data is permanently deleted from our infrastructure including MinIO object storage and PostgreSQL. GraphRAG entities associated with your organization are purged. We do not retain any customer data after the 30-day grace period.

NexusQA integrates with GitHub via webhook receivers for pull_request.opened and pull_request.closed events. When a PR is opened, the qa_pr_test_chain skill (Tier 3) triggers automatically — running regression tests, visual diffs, performance checks, and posting results as a PR comment with check status. When the autonomous pipeline creates a fix, it opens a PR with the label "qa-auto-fix" and assigns reviewers from your CODEOWNERS file. The Playwright test workers run in dedicated Kubernetes pods that auto-scale based on queue depth.

NexusQA runs on the Nexus Backplane which provides: AES-256 encryption at rest for all credentials and API keys, JWT-based authentication with organization-scoped Row Level Security on all database tables, Istio service mesh with mTLS for all inter-service communication, NetworkPolicy restricting ingress to gateway and nexus-alive only, and GDPR-compliant data handling with configurable data residency. The platform is open-source under Apache 2.0, which means you can audit every line of code. We do not yet have SOC 2 Type II certification — we are transparent about this rather than implying it.

We offer a 14-day free trial on Professional (no credit card required). We do not offer a permanent free tier because NexusQA requires significant infrastructure — Playwright workers running 3 browser engines, PostgreSQL with partitioned tables, GraphRAG (PostgreSQL + Qdrant + Neo4j), MinIO object storage, and BullMQ job queues. These have real infrastructure costs. What we promise instead: if you try it and it does not outperform your current QA stack, you pay nothing for those 14 days.