A self-hosted AI-powered analytics framework that connects test results with quality insights, risk analysis, and release decisions.
What Can QA Data Reveal About Customer and Revenue Risk?
QA data contains signals that can point to customer and revenue risk, but those signals are often buried in test reports, failure logs, and execution trends. A failed payment test may indicate a broken revenue path, while recurring checkout or authentication failures can point to friction in critical customer journeys. The challenge is not simply knowing which tests failed, but understanding what those failures could mean for customers, revenue, and release decisions.
What Is TR-Insight: An AI-Powered Test Analytics Framework?
TR-Insight by Testrig is a self-hosted, full-stack AI-powered test analytics framework that brings test results from Allure, spreadsheets, files, databases, APIs, and live Playwright or Cypress runs into a common analytical schema.
It combines a role-aware dashboard, conversational AI, and deterministic chart generation to help teams ask plain-English questions, analyze test results through data and visualizations, and make evidence-based release decisions without changing their existing testing setup.
TR-Insight runs within your environment, keeping sensitive data secure, with single-command deployment and support for existing test results.
TR-Insight: Key Capabilities
Natural Language Test Analytics
Ask questions about test results in plain English, such as “Which module has the most failures this sprint?” Multi-part questions are handled as separate queries, with aggregate results calculated across the complete dataset.
AI Test Reporting Dashboard
Explore test results through deterministic charts, tables, and automatically generated insights. The chart engine produces repeatable and auditable visualizations, while AI generates the underlying query rather than chart code.
Unified Test Data Analytics
Connect Allure reports, CSV, Excel, JSON, and Parquet files, MySQL and PostgreSQL databases, and REST APIs through a common analytical schema. Data can be ingested through a drop-box folder; CI upload API, dashboard, or CLI.
Live Test Execution Analytics
Monitor test executions as they run and bring completed results into historical analysis. Live reporting supports Playwright and Cypress executions, is designed not to fail the underlying test run, and supports sharded parallel runs.
Defect and Test Trend Analysis
Compare builds to identify regressions, recurring failures, and quality drift across releases. Build-over-build analysis provides a clearer view of changing application quality.
Role-Based QA Analytics
Provide relevant test insights for CTOs, QA managers, and SDETs while controlling access through account roles and team-specific workspaces. Permissions are enforced server-side.
Private AI Test Analytics
Connect Gemini, OpenAI, Anthropic, Groq, Ollama, or any LiteLLM-supported provider. Administrators can switch providers at runtime, test connections without restarting, or use local AI models for greater control over sensitive test data.
Self-Hosted Test Analytics
Run the framework within your own environment without an external database or SaaS dependency. Security controls include JWT authentication, bcrypt hashing, forced first-login password changes, append-only audit logs, and token-spend quotas to control AI usage and costs.
How TR-Insight Turns Test Results into Actionable Insights
- Automated Test Reporting That Saves Hours Every Week – No more exporting, cleaning, and summarizing reports by hand. Questions that used to take an afternoon are answered while the meeting is still running.
- Evidence-Based Go/No-Go Release Decisions– Every answer is backed by the data behind it, so go/no-go decisions rest on evidence instead of opinion.
- Detect Regressions and Flaky Tests Before Release – Regressions and unstable areas surface while a fix is still cheap, not after they’ve already cost you a release.
- Predictable AI Costs with Per-User Token Limits – Token limits are set per user and checked before every AI request. AI spend becomes a known line item, not an open-ended bill.
From Test Results to Decisions
See how your existing test data can support faster, evidence-based release decisions.
Talk To Our Experts
Let Us Take The Wheel: Entrust Your Testing To Our QA Professionals
