Repository navigation
Compared With Other Tools
Testosterone keeps the history of your test runs and tells you what changed: what broke, what's flaky and what got slower. It does that for any runner that writes JUnit XML (or that it can run itself), as a pip install on a laptop or one container for a team, with no account and no per-seat bill. This page says where that fits next to the tools teams already know, and when one of them is the better choice.
Features and prices of the other tools change often. This page describes them as of October 2026, from their own docs; check theirs before you decide.
| Testosterone | Allure Report | ReportPortal | Currents | Trunk Flaky Tests | Datadog Test Optimization | |
|---|---|---|---|---|---|---|
| Hosting | Your machine or your server | Static files | Your server (several services) or their cloud | Their cloud | Their cloud | Their cloud |
| Price | Free, MIT | Free | Free self-hosted; paid cloud | Paid, by usage | Free tier, then paid | Paid, part of Datadog |
| Runners | Any that writes JUnit XML, plus pytest, Behave, BehaveX and Playwright adapters | Most, through Allure adapters | Most, through agents | Playwright, Cypress | Most, through JUnit upload | Most, through Datadog libraries |
| History across runs | Yes, with per-test diffs between any two runs | One run, with an optional history folder | Yes | Yes | Yes | Yes |
| Flaky test detection | Yes, plus quarantine in config | Retries only | Yes | Yes | Yes | Yes |
| Per-test history page | Yes: outcomes, flake rate over time, duration, owner | Per run | Yes | Yes | Yes | Yes |
| Traces, screenshots and videos | Playwright attachments on the test page | Yes, the richest per-run view | Yes | Yes, streamed live | Not its focus | Through Datadog APM |
| Runs the tests for you | Yes (testo run, cycles of stages) |
No | No | Orchestrates Playwright and Cypress | No | No |
| Test case management and manual testing | No | No (Allure TestOps does) | No | No | No | No |
- You want the richest report for a single run. Allure Report's per-run view (steps, retries, attachments) is deeper than Testosterone's. They work together: Testosterone can write Allure reports as one of its reporters and link to them from each run.
- You need test case management or manual testing. That's Allure TestOps, Testmo and similar tools. Testosterone only knows about tests that ran.
- You want failures sorted automatically, with ML, across a large organisation. ReportPortal's auto-analysis learns from past decisions. Testosterone remembers a cause someone marked and applies it to failures with the same message, which is simpler and easier to predict, but it doesn't guess.
- You run Playwright or Cypress at scale and want CI load balancing. Currents splits and orders specs across machines and streams results live. Testosterone records what ran; it doesn't schedule specs across CI machines.
- Your team already lives in Datadog. Datadog Test Optimization links tests to traces, deploys and test impact analysis. Testosterone has no equivalent.
- Test data can't leave your network, or you don't want another SaaS and another per-seat bill.
- You use more than one runner (pytest, Playwright, Behave, a shell script) and want them in one history.
- You want "what changed since
main" on every pull request without wiring up a hosted service: one CI step with the GitHub Action, orpytest --testoon your own machine. - You want to try it in minutes:
pip install testo-core,testo ingest results.xml,testo ui. The Getting Started guide has the details, and the live demo's try-it page reads a JUnit file in your browser with nothing installed.
Generated from docs/Compared With Other Tools.md. Edit it there; changes made in the wiki are overwritten.
Architecture
CLI Commands
Processes & Guides
- CI-CD Pipeline Setup
- E2E Harness Operations Guide
- GitLab Pages Demo
- Publishing Docker Images
- Publishing to Artifactory
- Publishing to PyPI
- Shared Resource Pools
- Team Server
Prompts & Snippets
Roadmap & Strategy
Specs & ADRs
- Alerts Where Developers Look - 2026-10-22
- Architecture Consolidation - 2026-10-06
- BehaveX Reports - 2026-10-21
- Command Adapter and JUnit Import - 2026-09-25
- Config JSON Schema - 2026-10-22
- Delta Comparison Policy
- Docs Site - 2026-10-19
- Docs Vault and Wiki Sync - 2026-10-07
- Flaky Test Quarantine - 2026-10-22
- GitHub Action PR Comment - 2026-10-19
- Ingest and Team Server - 2026-10-22
- Interview Readiness Fixes - 2026-10-07
- Library Packaging
- Open Source Features - 2026-10-19
- Phase 3 Delta Comparison Engine Plan
- Phase 3 Unified Dashboard Plan
- Phase 4 BYOK and Failure Analysis
- Phase 5 UI Redesign - Cycles-First Navigation
- Playwright Adapter - 2026-10-19
- Playwright JSON Import - 2026-10-22
- Plugin API - 2026-10-19
- Product Roadmap Work - 2026-10-25
- Product to Ten - 2026-10-25
- Project Health - 2026-10-26
- Specs & ADRs
- Repository Pattern - Database-Agnostic Refactor
- Resource Pools and Leases - 2026-10-26
- Runner Coverage - 2026-10-26
- Scale Runs and Stages - 2026-10-24
- Switchable Team Features - 2026-10-24
- Throwaway Stage Containers - 2026-10-24
- UI Restyle - 2026-10-19
Testing Workflows
Vault