Get started in two commands

Quickstart

Install the agent, point it at your app, and read the report. First live result in about 5 minutes.

Install susatest-agent from PyPI, then run a test against a web URL or an APK. SUSA does the rest — it explores the running app like a real user and writes the findings to a report. It never needs your source, repo, or secrets.

terminal

# 1 · install the agent
$ pip install susatest-agent

# 2 · point it at a web app …
$ susatest-agent test https://app.com

# … or an Android APK
$ susatest-agent test app.apk

→ exploring autonomously · no script written
→ first live result in ~5 min · report ready

Black-box. SUSA tests the running app the way a user would. No instrumentation, no code access.

02

Four ways to run

Run anywhere

The same engine and the same report, wherever the app runs — our cloud, a local emulator, your own device, or a managed device lab.

Default

Cloud + Proxy · zero setup

Cloud + Proxy

Paste a URL and go. Runs on our cloud through residential proxies with country targeting and full browser stealth.

susatest-agent test https://app.com

Local · no hardware

Android emulator

No phone needed. SUSA boots a local Android emulator, installs your APK, and explores it over ADB.

susatest-agent test app.apk --emulator

Max realism

Your own device

Point SUSA at a phone or browser you own — your real IP and device fingerprint. Android via ADB, iPhone via USB (no Mac needed), or any local browser.

susatest-agent connect --web

Real hardware

Device labs

Pass-through to BrowserStack, AWS Device Farm, or Sauce Labs for real Android & iOS devices on demand.

susatest-agent test app.apk --lab

03

Ship it in your pipeline

CI & regression

Run SUSA on every release. It exits non-zero on failure and writes a re-runnable regression suite for you.

Non-zero exit on failure — fail the build when a critical flow breaks.
JUnit XML output — drops straight into your existing test reporting.
Works with GitHub Actions, GitLab CI & Jenkins — one command, no plugins.
Auto-generated regression scripts — Appium for Android, Playwright for web, written from the run.

.github/workflows/qa.yml

# run on every push — fail the build on a regression
$ susatest-agent test https://staging.app.com --junit results.xml

→ functional · a11y · security · pentest · perf
→ JUnit written · Appium + Playwright suite generated
→ exit 1 on failure

04

Drive it from your agent

MCP & agents

SUSA exposes an MCP server — kick off a run and pull the report without leaving your editor.

Connect from Claude Code, Cursor, or any MCP-capable agent. Your agent can start a test, watch it run, and read back the findings — so you can go from bug to fix in one place.

Start a run against a URL or APK straight from chat.
Pull the report — findings, repro steps, verdicts — back into the agent.
Works with any MCP client; no SUSA-specific glue code.

05

Nine disciplines, one pass

What it tests

Every run covers fourteen kinds of testing in parallel — driven by real persona exploration, not a script you wrote.

01 / functional

Functional

Every flow proved end-to-end with PASS/FAIL verdicts and repro steps.

02 / a11y

Accessibility

WCAG audit — contrast, touch targets, focus order, screen-reader, text scaling.

03 / security

Security

OWASP Top 10 — secret leaks, insecure storage, debuggable builds, auth bypass.

04 / pentest

Penetration

Active exploitation via the Mole engine — SQLi, XSS, IDOR, JWT, BOLA, SSRF.

05 / api

API contract

Capture endpoints, validate contracts, detect schema drift across builds.

06 / perf

Stress & performance

Rapid tap, rotation, slow network, low memory, FPS and janky-frame monitoring.

07 / visual

Visual regression

Per-build pixel + structural diff so layout breaks are caught before users see them.

08 / usability

Usability

UX friction, dead buttons, navigation loops and abandonment paths.

09 / llm

LLM observability

Trace LLM calls; judge evals catch hallucination, drift and safety regressions.

Docs