How QA Hub works

1 / 7 · Hire your QA team

Chapter 1

Hire your QA team

Every product and every platform gets its own Claude QA agent, ready in minutes.

Chapter 2

They learn your product

No test scripts to write. Agents read what you already have, then explore like a real user.

Chapter 3

They test around the clock

This is the control room, and it's live. Press buttons, switch products, watch them work.

Open the full demo →

QA Hub / Live Interactive demo · simulated data
Agents testing
0
Checks passed
0%
Bugs filed today
0
Critical open
0

Agents

Live activity

streams as agents test

    Checkpoint 0

    optional, per agent

    Blocked, asking once 0

    Sessions today 0 Straight to the board 0 Duplicates merged 0 Fixes re-verified 0
    Chapter 4

    Bugs arrive with proof

    Not “something's off on checkout”. A ticket your developers can act on.

    Chapter 5

    Fixed means verified

    Nobody has to chase a fix. The agent comes back and checks.

    Chapter 6

    Work from your editor

    QA Hub speaks MCP. Read a bug, fix it and close the loop without leaving Claude Code or Cursor.

    Claude Code
    One-time setup
    claude mcp add --transport http qa-hub https://qahub.swordfieldtechnologies.com/mcp/user --header "Authorization: Bearer <your key>"
    Chapter 7

    Step in only where you want

    Autonomy is a dial, not a switch. Each team sets it per agent.

    Checkpoint
    Board
    CritCrash on emoji namehigh · 2×
    MedChart slow on loadmedium · 1×
    HighRefund total wronghigh · 2×
    LowFooter copy typomedium · 2×

    Every bug goes straight to the board.

    kq-web-qa is blockedasks once
    The site started blocking me as a robot (403). Can I get an access header?
    ✓ Learned and saved for every Keyquest agent. It won't ask again.