ParalleX explores AI and applied data systems. Focus areas include bounded AI workflows, evaluation methods, data analysis and research software. Published repositories should identify whether their results are demonstrations or production-ready tools.

We build in the open. ParalleX Labs Inc. publishes its tools, methods and learning materials under open licences (usually Apache-2.0 for code and CC BY 4.0 for written material), so public-interest teams can use them, check how they work and adapt them freely. Open work is easier to trust, because anyone can see exactly how a result is produced.

Our own work, and only ours. Everything we publish here is our own, built from public data and synthetic examples. It never includes client data, client projects, or anyone else's confidential information or intellectual property.

• OpenData Sentinel: quality checks for municipal open data, with an accessible report and a correction backlog.
• Glossary Kit: glossary-as-code for public-sector data governance teams.
• Surrey ArcGIS public-data proof: City of Surrey development applications as schematic 3D extrusions with the ArcGIS Maps SDK for JavaScript, built from public data with an accessible facts list. Not a City system.
• civic-publish-kit: a public-data rehearsal of a reviewable update cycle for a City open-data layer, with provenance, plain-language validation reports and local rollback.
• Massing Provenance Kit: schematic building massing (GeoJSON and glTF) where every height carries its stated source, uncertainty and validation report.
• Humanitarian AI Risk Screen: a seven-step screen and training materials that help humanitarian and NGO staff decide whether and how to use an AI tool.
• Humanitarian AI Training Kit: open, bilingual (English and French) and accessible training on responsible AI use in humanitarian action, with a live site and tagged PDFs.
• AI Transparency Card: a bilingual (English and French) tool, live in the browser, for building a one-page transparency card for one proposed AI use, using SAFE AI's published tiers. Nothing you enter leaves the browser.
• Before You Paste: a bilingual practice lab, live in the browser, for checking text for personal and sensitive details before it goes into a generative AI tool. It runs on the device, asks permission questions first, and sends nothing to a server.
• Source Check Lab: a bilingual practice lab, live in the browser, for checking an AI-written summary against its sources, claim by claim. People commit a verdict first; optional hints run on the device.
• Translation Check Lab: a bilingual practice lab, live in the browser, for checking short humanitarian broadcast SMS messages and their translations before sending. It flags changed numbers, times, contacts, negation and dropped instructions, and a person's review record gates the final text; optional translation models run on the device.

We are open-sourcing more of our tools and learning materials in stages. Each new repository appears here after its tests, independent review and accessibility checks.

Contact: info@parallexlabs.ca | Website: https://parallexlabs.ca/