001
No commands
Attach your files or name a Jira issue in Claude Code, claude.ai, Cowork, or the desktop app and say what you need. No setup, no configuration.
CaselyAI QA test case generator
For QA engineers with a spec in a PDF and TestRail columns to match. Approve one test plan, get back cases in your team's format, a Postman collection for API specs, and a list of spec gaps. Free, open-source.
410 installs from Anthropic's plugin directory · 22,467 git clones on GitHub · 85/100 on a benchmark spec with five planted defects
all_test_cases.xlsx
7 casesIDTitleStepsExpected result
AUTH-014Password reset link expires after 30 minutesRequest reset, wait 31 min, open linkLink is rejected, prompts a new request
AUTH-015Reset link is rejected on second useUse reset link, reuse the same linkSecond attempt is rejected
BILL-002Card declined during plan upgrade keeps old planTrigger upgrade, decline card at checkoutAccount stays on the prior plan
BILL-007Invoice total rounds to two decimals at 19.995Generate invoice with a 19.995 line itemTotal displays as 20.00
IMP-021XLSX import with a renamed header columnRename a required header, import fileImport fails with a clear column error
IMP-022Import of 5000 rows completes in one passImport a 5000-row XLSX fileAll rows import without timeout
SES-009Session survives a token refresh mid-requestStart request, force token refresh mid-flightRequest completes on the new token
Why not just prompt a chat?
A chat prompt writes cases. Casely writes them in your format, checks the spec first, and hands back a file your tracker imports.
| What you need | Prompt in a chat | Casely |
|---|---|---|
| Your team's columns, order, and ID scheme | You describe them each time | Copied from one example, saved |
| Gaps and contradictions in the spec | Only if you ask | Listed in the plan, before any case |
| Your OK on coverage before writing | No | One pause, at the plan |
| One file TestRail imports in one pass | You format it | all_test_cases.xlsx or CSV |
| Runnable Postman collection for API specs | No | Yes, tokens as variables |
| Arabic, Korean, and other non-Latin text in the export | Unchecked | Round-trip tested, cell for cell |
| Bug report from a trace ID, data masked | You paste and redact the logs | /casely:bug in your format |
| Jira issues and Confluence pages | You copy and paste them | Read-only connector |
Comparing paid test case tools? Run this checklist before you buy.
Attach and scope.
Claude reads the attachments itself, PDF, DOCX, XLSX, TXT, MD, or pasted text, with no parser and no OCR setup. Name a Jira issue, an epic, or a Confluence page instead, and Casely reads it through a read-only Atlassian connector. Add one file of test cases your team already wrote, and Casely copies its columns, their order, the tone, and the ID scheme.
- No parser, no OCR setup, no config file
- Copies your team's columns and ID scheme, saved once
- PDF, DOCX, XLSX, TXT, MD, Jira, or Confluence
wallet-withdrawal
requirements_wallet.docxdocx
example_test_cases.xlsxxlsx
Here are the requirements for the Wallet module and example test cases my team writes. I need test cases for the withdrawal flow.
Viewed 2 files, ran 2 commands
Reading both files now. I'll match the columns and ID scheme from your example test cases before scoping the withdrawal flow.
Scoping withdrawal flow
Intake complete · 18 requirements · team format matched
Approve the plan.
Casely proposes coverage per module, estimates how many cases that is, and names what the spec gets wrong before writing anything. If the requirements describe an API, it says here how many cases are API-level and that those will also come back as a Postman collection. Then it stops and waits for your approval, the one pause in the whole run.
- Coverage estimate per module, before writing code
- Untestable wording and gaps named up front
- Says up front which cases are API-level and runnable
- One pause in the whole run, then Casely writes the rest
wallet-withdrawal
Here's the coverage plan and the gaps I found. Want me to go ahead and write the cases?
Thought for 4s
I've analyzed the withdrawal requirements and found these key areas. Does this plan look right before I write the cases?
ModuleEst.Notes
Card Binding3Card ≥3 days old; no linked card
Amount Validation4Min/max limit, negative
Daily Limit3Within limit, exceeds 150k
Get one Excel back.
Casely exports every case into a single all_test_cases.xlsx, or a UTF-8 CSV if your tracker wants one, plus one markdown file per case. TestRail, Qase, Zephyr, and Xray each import that file in one pass. API cases also come back as a Postman collection with an environment file and run instructions. In Claude Code, /casely:run then walks the UI cases in a browser.
- Equivalence, boundaries, decision tables, error guessing
- Single all_test_cases.xlsx, or a UTF-8 CSV
- Imports into TestRail, Qase, Zephyr, and Xray
- API cases also come back as a Postman collection
- /casely:run walks the UI cases in a browser
wallet-withdrawal
Approved. Go ahead and export the full suite.
Thought for 6s
Cases are written and exported. Equivalence, boundaries, decision tables, and error guessing are all covered. The 12 API cases are in a Postman collection too. Fill in baseUrl and authToken, then run.
all_test_cases.xlsx
Export complete · 21 cases · TestRail-ready
casely_api_collection.postman_collection.json
12 requests · 4 variables · Postman & Newman
Find the bugs. Casely writes the test cases.
Free and open source under the MIT license. Nothing installs on your machine: the export step runs on tools that already ship inside Claude's code execution environment.
002
Your columns
Attach one file of test cases your team already wrote. Casely copies the headers, their order, the tone, and the ID scheme, and saves them so you attach it only once.
003
Reports the gaps
Untestable wording, contradictions between sections, and error paths nobody wrote down: all named in the plan, before Casely writes a case.
004
One file to import
A single all_test_cases.xlsx holding every case under your own headers, or a UTF-8 CSV if your tracker wants one. Import it once instead of forty times.
005
API cases you can run
Endpoints in the spec? The API cases also come back as a Postman collection, with the base URL and tokens already pulled out into variables.
006
Runs them, writes the bugs
In Claude Code, /casely:run walks the UI cases through Playwright and marks each pass, fail, or blocked. /casely:bug turns a failure or a trace ID into a bug report in your team's format.
You approve before Casely writes.
Casely stops at the plan and tells you what the spec is missing. The coverage decisions stay with you.
Want it in the browser for your team, with file uploads and review? Join early access.
No. Writing test cases needs no command and no setup step. Attach your files or name a Jira issue in Claude Code, Codex, claude.ai, Cowork, or the Claude desktop app, say what you need in plain language, and answer when Casely asks which modules to cover. It pauses once, at the test plan. Two optional commands exist in Claude Code: /casely:run walks the finished cases in a browser, and /casely:bug writes a bug report.