Inflectra Internal Tool

Spira Project
Setup Tool

Automated Spira environment provisioning
via the REST API — driven by a JSON config file

DERMOT CANNIFFE  ·  MAY 2026

Today's agenda

  • 1 The problem we're solving
  • 2 What the tool does
  • 3 How it works
  • 4 The AI angle
  • 5 Questions

The problem

Every new Spira engagement starts the same way.

  • SE or PS engineer logs into Spira admin UI
  • Manually creates programs, products, releases
  • Manually configures custom fields — one by one
  • Manually creates test folders and populates test cases
  • Repeats this for every client, every environment, every phase
This is slow, error-prone, and impossible to version-control.

At scale it gets worse

What we want

  • Consistent environments across clients
  • Reproducible setups for demos and PoCs
  • Fast onboarding for new engagements
  • A record of what was provisioned

What we get

  • Inconsistent naming conventions
  • Missing custom fields
  • No audit trail
  • Knowledge locked in one person's head

The solution

A config-driven Python script that provisions a complete Spira project structure via the REST API v7.

Define your structure in a JSON file. Run one command. Done.
  • Creates products, releases, custom fields, folders, and test cases
  • Fully idempotent — safe to re-run, skips what already exists
  • One JSON file per client — version-controlled, shareable, auditable
  • Multiple structure files supported — one repo, many clients

The structure file

A single JSON file describes everything to be created.


{
  "program": {
    "name": "Oetker Portfolio",
    "products": [{
      "name": "Oetker Boost phase A",
      "releases": [
        { "name": "Oetker Boost phase A - Release 1" },
        { "name": "Oetker Boost phase A - Release 2" }
      ],
      "customFields": {
        "testCases": [{ "name": "Transaction Code",
                        "type": "list", "values": ["VA01","ZVA01"] }],
        "testSets":  [{ "name": "Country",
                        "type": "list", "values": ["DE","IT","FR"] }]
      },
      "testFolders": [{
        "name": "Oetker SD",
        "testCases": [
          { "name": "Create Standard Sales Order",
            "description": "Verify VA01 creates a sales order." }
        ]
      }]
    }]
  }
}

How it works

1
Load credentials from .env and validate connection to Spira
2
Read the structure JSON file
3
Verify the program exists in Spira (must be created manually — API limitation)
4
For each product: create it, resolve its template, then create releases, custom fields, folders, and test cases
5
Print a full summary of everything created
Each step checks for existing resources first — re-running is always safe.

Code architecture

Thin client + service layer + runner. Each layer has one job.


spira_setup/
├── client.py       ← all HTTP: auth, retries, error handling
├── runner.py       ← reads JSON, orchestrates everything
└── services/
    ├── projects.py    ← create/find products & programs
    ├── releases.py    ← create releases
    ├── templates.py   ← custom lists & custom properties
    └── test_cases.py  ← folders & test cases
  • Adding a new artifact type = one new service file
  • API version change = one file to update (client.py)
  • New client = one new JSON file, zero code changes

The AI angle New

The structure file has a formal JSON Schema. Any LLM can generate a valid file from a plain-English brief.

"Generate a Spira structure file for a logistics client. Two products — WMS and TMS. Releases for Q3 and Q4 2025. Test cases covering order creation, shipment tracking, and returns. Add a Country field (DE, IT, FR) to test sets."
  • Works with ChatGPT, Claude, Kiro, Copilot — any LLM
  • Schema enforces correctness: required fields, valid types, no extra keys
  • SE or AM describes the client's needs → LLM produces a ready-to-run file
  • VS Code autocomplete and validation included for free

End-to-end workflow

SE / AM
Describes client needs to an LLM
→
LLM
Generates structure.json from schema
→
SE / AM
Reviews structure file
→
Script
python setup.py
→
Spira
Fully provisioned in seconds
From brief to live Spira environment — without touching the admin UI.

Questions?

⌥ Repo github.com/dermotcanniffe/SpiraProvisioner
▶ Deck dermotcanniffe.github.io/SpiraProvisioner

BUILT WITH PYTHON · SPIRA REST API V7 · REVEAL.JS