Cognitia Labs is an independent lab for research-oriented software, open-source experiments, and practical AI systems, focused on inspectable tools at the intersection of applied AI, software infrastructure, digital trust, and access to information.
Highlighted projects
Digital signatures · PDF · PKCS#11
A command-line tool for signing PDFs locally with Uruguay's electronic ID, via PKCS#11. Produces CMS/PAdES signatures that validate in standard PDF tooling. The project explores how national digital identity infrastructure can interoperate with local, scriptable document workflows.
Not affiliated with AGESIC or any Uruguayan government agency.
View on GitHubOpenPGP · eID · Cryptographic identity
A public binding between my OpenPGP key and my Uruguayan eID signature. The project explores how open-source cryptographic identity and legally recognized digital identity can be connected through independently verifiable attestations.
View on GitHubPython · LLM pipelines · HTML
A Python library for splitting HTML into chunks that preserve DOM structure, designed for LLM pipelines that need to process long web pages. It builds a DOM tree and selects content-rich regions of interest, making the resulting chunks more meaningful than plain text splits.
View on GitHubPython · Serialization · Tooling
A small Python utility for turning complex Python objects into JSON-ready structures, with support for Pydantic models, dataclasses, and circular references. Originally built as infrastructure for logging and inspecting LLM tool calls.
View on GitHubPosition
AI is still entering culture, and we are still learning which parts of it should be handled by models, and which should remain ordinary software. The work I find meaningful sits at that boundary, in systems dealing with fuzzy inputs, imperfect information, human judgment, and situations where approximate reasoning can be useful, but only when its limits are explicit.
Cognitia Labs explores that boundary through small, inspectable software artifacts, including tools, experiments, and open-source infrastructure designed to make AI systems easier to understand, operate, and question. The cases that draw me in most are usually the ones that lead into questions of trust, access to information, cryptographic identity, and the role software plays in public life.
About
Cognitia Labs is run by me, Carlos A. Planchón, a software developer based in Uruguay. This is where I give shape to some of the questions and unresolved ideas that keep coming back, tinkering with them until they become software.
Other writing and contact details live at carlosplanchon.com.
Contact
For conversations about applied AI, open-source tooling, digital trust, or software infrastructure.
