Media & Research
Explore Eka.care’s published work across medical AI evaluation, multilingual healthcare models, speech recognition, and medical document understanding, with a strong focus on the Indian healthcare context.Introducing KARMA – OpenMedEvalKit
Read more An open-source framework for standardized, transparent evaluation of medical AI and LLMs.Designed to enable reproducible benchmarking and responsible medical AI development.
Parrotlet-E & Eka IndicMTEB
Read more Multilingual models and benchmarks built to address India’s diverse healthcare language needs.Focused on improving medical understanding and evaluation across Indian languages.
Beyond Traditional WER: Semantic WER for ASR
Read more Explains why traditional Word Error Rate is insufficient for medical ASR systems.Introduces Semantic WER to better measure clinical correctness in Indian healthcare speech.
Parrotlet-V Lite (4B) – Vision LLM
Read more A purpose-built vision-language model designed for parsing medical documents and records.Optimized for extracting structured data from real-world clinical artifacts.
Parrotlet-A EN (5B) – Medical ASR LLM
Read more A medical-grade speech recognition model tailored for English ASR in Indian healthcare.Built to handle clinical vocabulary, accents, and real consultation audio.
Advancing Healthcare AI Evaluation in India
Read more Release of four healthcare-focused evaluation datasets for AI benchmarking.Supports better measurement, comparison, and progress tracking of medical AI systems.