Reflex, by BQA Solutions
Know your software works. Every morning.
Reflex clicks through your applications like a user, checks your integrations and takes over recurring manual work. Overnight, or whenever you like. When something breaks you see exactly where. And your AI helps fix it.
- Hosted by us or on your own servers
- Runs inside your network
- A demo on your own application
Built in
- Playwright
- Selenium (Chrome, Firefox)
- REST
- SOAP
- JSON, XML and YAML
- SQL databases (JDBC)
- CouchDB
- Excel
- CSV
- PDFs and scans (OCR)
- ZIP, TAR and 7z
- SFTP
- SSH
- Windows shares (SMB)
- SharePoint
- Send and read email
- Telegram
Manual work takes time. And things still break.
Today
Regression tests take days
Before every release someone clicks through the same screens. Every release waits for it.
With Reflex
Everything tested overnight. In the morning you know whether you can release.
Today
Routine work depends on people
Excel reports, file transfers, checking mailboxes. When that one colleague is out, it waits.
With Reflex
It runs on schedule, also when that colleague is off. Every step is recorded.
Today
You find out too late
A broken integration shows up when a customer calls.
With Reflex
When something fails you are notified right away. With the step where it went wrong.
How it works
-
16:00 During the day
Describe what should happen
Your tester or your AI assistant writes a short script: sign in, fill in a form, check the result. Or: fetch an Excel file, process it and mail it.
tests/login.lua -- Can a customer get a quote? local browser = Play.start() browser.navigate("https://www.verzekeraar.example/auto") browser.locator("id=kenteken").fill("AB-123-C") browser.locator("id=postcode").fill("3833 LD") browser.locator('role=button[name="Get a quote"]').click() Assert.contains(browser.locator("css=.premie").text(), "per month") -
02:00 At night
Schedule it
Every night, every 15 minutes or on every release. Reflex runs it on an agent inside your network.
-
07:00 In the morning
See the result, fix it fast
Follow every step live and replay it later. When something fails you see exactly where. AI proposes a fix, you approve it.
One platform for testing and recurring work
Test automation
Regression and smoke tests for web applications and APIs, run in parallel, with a pass rate for every run.
- Browser tests with Playwright or Selenium
- API checks over REST and SOAP
- Started from your CI/CD pipeline
- Broken selectors proposed and verified
Process automation
Recurring work that is manual today, done automatically and on time.
- Excel reports, mailed at 06:30
- Files over SFTP, Windows shares and SharePoint
- Mailboxes read and answered
- Checks on databases and integrations
Works with the AI tools your team already uses
Your AI assistant writes scripts, runs them on your agents and finds out why a test failed. It acts with the user’s permissions, and a person approves changes.
- Claude Code
- Codex
- OpenCode
You Why did the quote test fail last night?
AI Step 11 could not find the ‘Get a quote’ button: it has a new id. I tried another selector and it works.
AI The fix is ready in Reflex and waiting for your approval.
Integrations are built in
Browsers, APIs, Excel, email, SFTP, SharePoint, databases and more. No plug-ins or extra licences.
Web applications
PlaywrightSelenium (Chrome, Firefox)
APIs and services
RESTSOAPJSON, XML and YAML
Databases
SQL databases (JDBC)CouchDB
Documents
ExcelCSVPDFs and scans (OCR)ZIP, TAR and 7z
Files and servers
SFTPSSHWindows shares (SMB)SharePoint
Messaging
Send and read emailTelegram
You stay in control
Runs in your network
Agents connect out to the server. You open no ports.
Roles and audit trail
Decide per project who can do what. Every change and action is recorded.
Encrypted secrets
Passwords are stored encrypted and never show in logs.
Single sign-on
Sign in with OpenID Connect or Microsoft Entra ID.
We host it, or you do
SaaS
BQA hosts and maintains Reflex. Your agents run inside your own network.
Self-hosted
Run Reflex on your own servers, with Docker, PostgreSQL and your own keys.
- Implementation and a first regression suite
- Training for testers and developers
- Maintenance, also temporarily
- Support remote and on site
Frequently asked questions
Do we need to be able to code?
A little helps. A Reflex script is short and readable, and your AI assistant writes most of it. We teach your team the rest.
Does it run inside our network?
The work happens on agents you place: in your network, on Linux or Windows. They connect out to the server, so you open no ports. We host the server for you, or you run it yourself.
Does it fit our CI/CD pipeline?
Yes. Your pipeline starts a test suite with one command and gets the result back, so a release stops when something fails.
Which AI tools work with Reflex?
Claude Code, Codex and OpenCode. They act with the user’s permissions, and every action lands in the audit trail.
What does Reflex cost?
It depends on your number of projects and agents, and on whether we host it or you do. We make you a tailored proposal after an introduction.
See Reflex work on your application
In 30 minutes we show you what Reflex does for your tests and processes.