Robot Framework Evidence Reporter

Business evidence reports in HTML, PDF and Word for each Robot Framework test case.

English · Español

User guide · Keyword reference · PyPI · Visual examples

Features

• Explicit page, element, desktop and existing-image captures.
• Optional milestones, business messages and configurable metadata.
• Offline HTML, PDF/Word output and light/dark report themes.
• Parallel recording, execution/rerun merging and a SHA-256 file manifest.

Installation

Python 3.12+ and Robot Framework 7. SeleniumLibrary is required only for browser captures and reuses its active browser.

pip install robotframework-evidence-reporter robotframework-seleniumlibrary

Quick start

*** Settings ***
Library    SeleniumLibrary
Library    EvidenceReporter
Test Teardown    Close All Browsers

*** Test Cases ***
Portal evidence
    Open Browser    https://example.com    headlesschrome
    Wait Until Element Is Visible    css:h1
    Element Text Should Be    css:h1    Example Domain
    Set Report Metadata    Application=Example    Environment=Demo
    Capture Page Evidence    Portal available    status=PASS

The included listener records JSON and images during execution. Generate the requested formats afterwards:

robot --outputdir results evidence.robot
rf-evidence build results/evidence --output reports --formats html pdf docx

Milestones are optional. Evidence without milestone_id belongs directly to the case.

Configuration and limitations

Normally omit output_dir: evidence uses Robot's ${OUTPUTDIR}/evidence, controlled by --outputdir. For an alternative location, use --variable EVIDENCE_OUTPUT_DIR:artifacts/evidence. An explicit import output_dir takes precedence over that variable, then the default. Configure strict on import when needed. Capture failures produce WARN by default without changing test status; strict=${True} requires a successful capture. Evidence accepts INFO/PASS/WARN/FAIL; SKIP belongs only to execution.

Use a new evidence directory per execution. Pabot stores evidence inside each worker’s output directory; after all workers finish, generate recursively with rf-evidence build results/pabot_results --output reports. Cases use independent UUID directories. Desktop capture requires a compatible graphical environment and captures the local process desktop, not a remote Selenium Grid desktop. Full-page capture is unsupported.

Images and notes are not automatically redacted. Usage errors and unexpected errors propagate. PDF/Word are static; HTML contains interactive controls. manifest.json is this package's inventory, not Allure results.

Examples

The Selenium example demonstrates business evidence recording. Public visual demos use clearly labeled fictitious data and illustrative images. See the manual for merge, image sizing, file inventories and capture errors.

Development and contribution

poetry install
poetry run ruff check src tests scripts
poetry run ruff format --check src tests scripts
poetry run python -m unittest discover -s tests -v
poetry run python scripts/validate.py
poetry run python docs/scripts/build_docs.py
poetry build

Submit changes through a pull request with passing checks. Update both documentation languages. Libdoc translations live in docs/translations/es/libdoc.json; builds reject missing or stale entries.

License

MIT. See LICENSE.