SPECTRA
AI-native test case generation and execution framework. From documentation to verified, grounded test cases — automatically.
Why SPECTRA?
| Manual approach | With SPECTRA |
|---|---|
| Write test cases in spreadsheets or Jira | AI generates from your docs |
| Test cases drift as docs change | Doc changes auto-flag outdated test cases |
| No systematic boundary/edge coverage | 6 ISTQB techniques applied automatically |
| Coverage gaps invisible until bugs | Three-dimensional coverage dashboard |
| Subjective test case reviews | Dual-model grounding verification |
| Locked into test management tools | Markdown files versioned in Git |
How it works
Your docs → spectra ai generate → AI analysis → critic verification → test-cases/*.md
Or just say it in Copilot Chat: “Generate test cases for checkout”
Key capabilities
- AI Test Case Generation — iterative sessions with dual-model verification
- Acceptance Criteria Extraction — MUST/SHOULD/MAY from docs with SHA-256 change tracking
- Three-Dimensional Coverage — docs × criteria × automation in one dashboard
- MCP Execution Engine — execute through Copilot Chat, Claude, or any MCP client
- Testimize Integration — optional algorithmic test data optimization (BVA, EP, pairwise)
- Git-native — Markdown + YAML, no database, no vendor lock-in
Quick install
dotnet tool install -g Spectra.CLI
spectra init
spectra ai generate
See Getting Started for prerequisites and authentication setup.