Unified Test Management
Powered by Any LLM

Getting Started

From install to your first gap report in under 10 minutes.

1. Install and initialise

   pip install testweavex
   tw init --llm-provider anthropic

   This creates testweavex.config.yaml in your project root and a testweavex/skills/ folder with the 10 built-in skill files.

2. Run your test suite

   tw                         # same as pytest — all flags work
   tw tests/login.feature    # run a specific feature
   tw -k smoke -n 4          # filter + parallel

   Results are stored automatically in .testweavex/results.db. No configuration required.

3. View your first gap report

   tw gaps --limit 20

   TestWeaveX compares your TCM against your automation suite and surfaces unautomated tests ranked by priority score. Generate automation for any gap with tw gaps --generate.

CLI Reference

Every pytest flag works with tw unchanged. TestWeaveX adds its own flags alongside.

CI/CD Example (GitHub Actions)

- name: Run tests
  run: |
    tw run --suite regression \
           --results-server ${{ secrets.TW_SERVER }} \
           --token ${{ secrets.TW_TOKEN }} \
           --sync-tcm testrail

Configuration

Create testweavex.config.yaml in your project root. All values support ${ENV_VAR} interpolation. Missing keys use defaults.

# testweavex.config.yaml
llm:
  provider: anthropic        # openai | anthropic | ollama | azure
  model: claude-sonnet-4-6
  api_key: ${ANTHROPIC_API_KEY}
  temperature: 0.3
  max_retries: 3
  timeout_seconds: 30

results_server: ${TESTWEAVEX_SERVER}  # optional — team mode

tcm:
  provider: none              # testrail | xray | none

gap_analysis:
  scoring_weights:
    priority:  0.30
    test_type: 0.25
    defects:   0.20
    frequency: 0.15
    staleness: 0.10
  match_threshold: 0.65
  top_gaps_default: 10

LLM Providers

Architecture

TestWeaveX is a pytest plugin with a thin CLI wrapper. All functionality flows through one of three pipelines.

Component Map

Stable ID Algorithm

Every test case gets a deterministic 64-character ID derived from its feature file path and scenario name. This ID is stable across machines, CI runs, and environments. The algorithm is frozen — never change it after first deployment.

import hashlib

def generate_stable_id(*parts: str) -> str:
    key = "|".join(parts).encode("utf-8")
    return hashlib.sha256(key).hexdigest()  # full 64 chars

# test_case_id = generate_stable_id(feature_path, scenario_name)
# feature_id   = generate_stable_id(feature_path)

Gap Priority Scoring

Gaps are ranked by a six-signal weighted score (0.0–1.0). Higher = automate first.

Contributing

Repository Structure

Adding a Custom Skill

The easiest way to contribute is a new skill YAML file. Create testweavex/skills/custom/your-skill.yaml:

name: custom/your-skill
display_name: Your Skill Name
description: What this skill generates

prompt_template: |
  You are a senior QA engineer.
  Feature: {feature_description}
  Generate {n_suggestions} test scenarios that...
  Return JSON: title, gherkin, confidence, rationale, suggested_tags

assertion_hints:
  - Verify primary outcome
tags: [custom]
priority: 3

Development Setup

git clone https://github.com/testweavex/testweavex
cd testweavex
pip install -e ".[dev]"
pytest tests/ -v