MCP-NATIVE TEST MANAGEMENT

The AI-driven testing platform your QA team actually runs on.

Caseract holds your products, test plans, cases and runs in one workspace, and connects that workspace to Claude Desktop over MCP, so an agent can draft cases, read your run history and tell you where coverage is thin.

Start with the free tier, unlimited test cases See how it works
NO CARD · UNLIMITED CASES ON FREE · AI ON PREMIUM
TRUSTED BY QA TEAMS AT
Zegates FILE HUB wad/crm Northwind HRM SIGNATRA
01 · What Caseract does

One library for every case. One place to run them.

Products hold test plans, plans hold cases, cases produce runs, runs produce reports. Nothing lives in a spreadsheet and nothing has to be re-explained to the agent. The same structure that organises your QA is the structure Claude reads.

CASE LIBRARY

Manual and automated, side by side

Priority, component, ticket link, steps and expected results on every case. Filter by any of them, then save the filter as a view.

EXECUTION

Triage a run in one hand

J / K to move, 1 2 3 to mark pass, fail or skip. Selection auto-advances and coverage updates as you go.

COVERAGE

Reports that answer one question

Per-product coverage, failing P1s, cases never run. Passed, failed, running and skipped are colour-coded everywhere they appear.

Inside the workspace

WAD CRM / TEST CASES MCP ✓ CONNECTED
Test cases
Your library of manual and automated cases.
PASSED
18
FAILED
3
NOT RUN
3
Search cases All priorities ⌄ All statuses ⌄ Component ⌄
ID
TITLE
PRIORITY
STATUS
TC-014
Guest can complete purchase
P1
PASSED
TC-015
Apply expired promo code
P2
RUNNING
TC-016
Tax recalculates on address change
P1
FAILED
TC-021
Signature merge keeps field order
P2
PASSED
TC-024
Bulk filter export respects tags
P3
SKIPPED
02 · MCP & AI analysis

Connect Claude Desktop to your test library.

Caseract exposes your workspace as an MCP server. Add it once in Claude Desktop and the agent can read products, plans, cases and run history, then write back drafted cases, tags and analysis. Everything it does lands in the activity feed, attributed and reversible.

01
Draft cases from a ticket. Paste an ID; the agent writes title, steps and expected results in your house style.
02
Ask about coverage. “Which P1s have no run this week?” answered against live data, not a stale export.
03
Analyse failures. The agent groups repeat failures, points at the likely component and proposes the missing case.
CLAUDE DESKTOP · MCP CONFIG READY
{
  "mcpServers": {
    "caseract": {
      "url": "https://mcp.caseract.com",
      "token": "csr_live_••••••••"
    }
  }
}
list_cases create_case get_run coverage_report analyse_failures
MCP access and AI analysis are part of the Premium plan.
03 · How to use it

From empty workspace to a green run, in six steps.

STEP 01

Create a product

Name the thing you're testing and invite your team. The product switcher at the top of the sidebar moves you between them.

STEP 02

Add a test plan

Group cases by release, regression suite or component. Plans can be enabled and disabled without deleting anything.

STEP 03

Write cases in the side sheet

Title, priority, component, ticket, then steps with their own expected result. The list stays visible while you write. ⌘↵ saves and opens a fresh sheet.

STEP 04

Or seed them with a ticket ID

Paste an ID into the AI seed row and press Enter. Drafted content is marked agent-authored until you confirm it. Premium.

STEP 05

Run and mark

Start a run, then work the list with the keyboard. Each case keeps its own history, so “last run” is always answerable.

STEP 06

Ask the agent what's thin

With Claude Desktop connected, ask for gaps, repeat failures or a summary for standup. Premium.

04 · Packages

The platform is free. The agent is Premium.

Unlimited test cases on the free tier, forever. Pay only when you want MCP access and AI analysis.

FREE
Testing platform
Everything a team needs to write, organise and run tests by hand.
$0 FOREVER
✓Unlimited test cases
✓Unlimited products, plans and runs
✓Up to 5 members
✓Keyboard triage, filters and saved views
✓Coverage dashboard and reports
✕No MCP or AI analysis
Start free
AI ENABLED
PREMIUM
MCP + AI analysis
The whole free platform, plus Claude Desktop connected to your library.
$15 / USER / MONTH
Billed annually: $180 per user per year, 20% off monthly.
✦MCP server for Claude Desktop
✦AI-drafted cases from a ticket ID
✦Failure clustering and gap analysis
✦In-app agent rail and command palette
✓Unlimited members
✓Everything in Free
Upgrade to Premium
ENTERPRISE
Premium at scale
For QA orgs with procurement, audit and self-hosting requirements.
Custom
✓SSO / SAML and audit log
✓Self-hosted MCP endpoint
✓Volume seat pricing
✓Onboarding and named support
Talk to us
Annual billing saves 20%. Change or cancel a seat any time. AI analysis uses your Caseract data only, never for model training.
05 · From QA teams

The side sheet alone changed how much we write down. Cases get authored during triage instead of after it.

SA
Sara Aziz
QA Lead, WAD CRM

I ask Claude which P1s haven't run this week and get a straight answer. That used to be a Friday spreadsheet.

DM
Daniel Mor
Engineering Manager, File Hub

We started on the free tier with 400 cases and never hit a wall. Premium was a decision about the agent, not about limits.

NK
Nadia Karim
Head of Quality, Northwind HRM
Start with the free tier. Unlimited test cases.
Add MCP and AI analysis whenever your team is ready.
Create a workspace Read the MCP setup