Papers
GitHub: https://github.com/eelab-dev/EEdigits
Peer-reviewed paper: https://arxiv.org/abs/2606.21451
Resources
PRAMANA: Proof-centric RTL Agentic Model for Assurance, Narrative, and Automation
PRAMANA—Proof-centric RTL Agentic Model for Assurance, Narrative, and Automation—is an open-source, AI-assisted platform designed to make formal hardware verification more intelligent, transparent and approachable. PRAMANA takes its name from the Sanskrit word meaning “proof,” “measure” and “means of knowledge". Verifying a semiconductor design is not simply about running a proof: engineers must also determine whether the verification properties are strong enough, choose an effective solver and understand why a proof has failed. PRAMANA brings these challenges together within a single quality-aware workflow.
The platform evaluates the strength of verification properties using controlled hardware mutations, analyses design characteristics to support solver selection, and converts complex counterexamples into clear, evidence-based explanations. This helps experienced verification engineers investigate results more efficiently, while also giving students and newcomers a practical route into formal verification.
PRAMANA is openly available on GitHub, with code, hardware examples and reproducible experiments that users can run, extend and contribute to. The research underpinning the platform has been peer-reviewed and accepted at SBCCI 2026, providing a strong academic foundation for its use in AI-assisted semiconductor design and verification.
