Applications · Prototypes · Research
Projects developed through
exploration and experimentation.
ParancU Lab brings together released software, experimental applications and ongoing research into document retrieval and evidence assessment.
Each project represents a different stage of development. Their capabilities and limitations are presented according to their current maturity.
ParancU for iOS
A released mobile application and a concrete result of the laboratory’s development work.
Import and prepare PDFs or text, including camera/image OCR. Multilingual embedding-based retrieval combines semantic matching with lexical signals, letting you inspect the retrieved source passages.
Retrieval runs locally after corpus preparation. Preparation currently sends source chunks to OpenAI using a user-supplied API key.
ParancU Agent Lab
A research prototype exploring browser-based retrieval, AI-assisted answers, source inspection and verification workflows.
Imported prepared corpora support browser-local retrieval only. Experimental server-assisted workflows send documents to the server and use OpenAI for preparation, answer generation and verification, with limited recovery.
Verification is an experimental, model-based safeguard, not proof of correctness.
ParancU 2.0
Investigating how to retrieve passages that contain the facts needed to answer a question.
The research explores factual representations, question-oriented retrieval, semantic relevance versus answer containment, alternative retrieval strategies and failure cases.
These are methods under investigation, not a validated or selected new retrieval architecture.
Explore ParancU 2.0 research →ParancU Decision Lab
Exploring whether retrieved evidence is sufficient to answer a question.
Investigations examine complementary evidence selection, missing-information diagnosis, verification boundaries and recovery strategies when retrieved evidence is insufficient.
Experimental findings do not establish general reliability. Decision Lab and ParancU 2.0 are not released products.
Explore Decision Lab research →Documentation & Resources
Elsewhere
Agent Lab on GitHub ↗ParancU on PeerPush ↗ParancU on TinyHunt ↗
Research evolves through what gets built and tested.
These projects reflect an ongoing process of implementation, experimentation and evaluation. They are not presented as a unified commercial platform or a completed product roadmap.