Spec2RunAI is an AI-native, no-code test automation tool from AgileAI Labs. It runs the manual test cases your team already has, written in plain English, against live web applications — as written. There is no command vocabulary to learn, no generated scripts, and no hard-coded selectors to maintain. Spec2RunAI is sold on its own and is also the test execution layer of Spec2TestAI, AgileAI Labs’ Quality Intelligence platform for the enterprise SDLC.
How it works: paste your test cases, or import them from Jira, Zephyr, qTest, Excel, or Word. An AI agent reads each step the way the tester wrote it — business jargon, multi-action steps, and expected results included — and resolves it against the live screen at run time. Steps that have proven themselves replay deterministically from selector memory with no model call, so repeat runs are fast and predictable. When the application changes, self-healing operates under Assertion Lock: Spec2RunAI may change how a step is performed, but never what it verifies. Every suggested rewrite is shown beside the original, and the stored test changes only when a person approves it.
Every run ends in a Run Evidence Report. It shows verification coverage per step (Verified, Not verified, or Failed, with the reason), failures classified by cause (application defect, test defect, environment, or data), a comparison with the previous run (Regressed, Fixed, New in suite, and a stability score per test case), and an AI provenance appendix that records what the AI did and did not change. Reports export to PDF with signature lines for QA review and business sign-off, and to Markdown for Jira, Slack, or pull requests.
Spec2RunAI runs packaged enterprise applications without object libraries, including Insurity, Guidewire, Salesforce, SAP, Workday, and ServiceNow, and has been tuned against live Insurity and Guidewire PolicyCenter environments. Tests run on headless Chromium in the cloud by default, or in real Google Chrome and Microsoft Edge on Windows and macOS through your own BrowserStack Automate account, with the same tests unchanged. Applications behind your firewall run through a local agent that connects outbound over HTTPS only, holds no AI provider key, and needs no inbound ports or VPN.
Also included: data-driven testing from the spreadsheet you already use, with one verdict and evidence set per row; two-factor sign-in for authenticator-app (TOTP) accounts from an encrypted secrets vault; autonomous exploration of unscripted flows inside engine-enforced guardrails, with a reproducing test case written for each finding; drift detection on passing steps every run; and a trigger, status, and callback API for GitHub Actions, GitLab CI, Jenkins, Azure DevOps, or any pipeline.
Enterprise access: single sign-on through Amazon Cognito, configured per customer, and role-based access control with roles including Super Administrator, Administrator, Tester, Business Analyst, Project Manager, Product Manager, and Developer. AgileAI Labs holds a SOC 2 Type II attestation, isolates data per tenant, and never accesses your source code.
Pricing is published: $3,000 per 100,000 AI engagements ($0.03 each). Four engagements on first resolution of a step, one per replay, and engagements expire after 12 months. Send us ten untouched manual test cases and we will run them and return the evidence report.
Who Is the Company Behind Spec2RunAI?
