Eka Care loves Open Source
31 projects, 1 upstream contribution, Updated 19 Oct 2026.
The code, models and datasets we build for Indian healthcare, open for anyone to use and improve.
Projects
-
Patra
Design a printable document, then render it with data.
npm install github:eka-care/patra#patra-engine-v0.1.0 -
Paged.js (Eka fork)
Table pagination fixes for Paged.js, used to print prescriptions and bills.
Fork of Paged.js 0.4.3
git clone -b eka/v0.4 https://github.com/eka-care/pagedjs -
Pagify SDK
Render HTML as paginated, print-ready PDFs in the browser, with headers and page numbers.
npm install @eka-care/pagify-sdk -
Echo SDK
One agent SDK for Bedrock, OpenAI and Anthropic, alone or inside CrewAI and LangGraph.
pip install git+https://github.com/eka-care/echo-sdk.git -
document-pii-redactor
Find and redact PII in document images and text. Built for Indian documents; runs on CPU.
pip install document-pii-redactor -
POS Terminal Mock
A stand-in card machine, so you can build a POS integration without the hardware.
npx github:eka-care/pos-terminal-mock -
Aangan
Compose a home screen from independent providers, ranked and capped per zone.
npm install github:eka-care/aangan#aangan-contract-v0.1.0 -
Awaaz
Collect product feedback - ideas, bugs, votes - and run it through a status lifecycle.
npm install github:eka-care/awaaz#awaaz-contract-v0.1.0 -
FHIR Prism
Read an ABDM FHIR bundle as a clinical document, in the browser.
-
KARMA OpenMedEvalKit
Evaluate medical AI models - text, image and audio - on one framework.
pip install karma-medeval -
Med Doc Classifier
Sorts a document image into 27 classes and flags medical and handwritten, in ~96M params.
hf download ekacare/med-doc-classifier -
Parrotlet-a 2.5 Pro
Medical speech recognition for Indian English, Hindi, Marathi, Kannada and Telugu.
hf download ekacare/parrotlet-a-2.5-pro -
Parrotlet-v 2.5 Pro
Reads Indian medical records - prescriptions, lab reports, invoices - into schema JSON.
hf download ekacare/parrotlet-v-2.5-pro -
Eka NFI MCQA
925 pharmacology questions drawn from the National Formulary of India.
hf download ekacare/Eka_NFI_MCQA --repo-type dataset -
Indian Drug MCQA
1,512 questions: name the generic composition behind an Indian drug brand.
hf download ekacare/indian_drug_mcqa --repo-type dataset -
Indian Protocols-Based Clinical Q&A
504 doctor queries graded against the clinical guideline they should be answered from.
hf download ekacare/indian_protocols_based_clinical_QnA --repo-type dataset -
Medical Calculator Evaluation
1,066 questions on 358 medical calculators: can a model do the maths without a tool?
hf download ekacare/medical_calculator_eval --repo-type dataset -
BODHI-M
A clinician-built graph linking 2,471 clinical concepts to their drugs and lab tests.
hf download ekacare/BODHI-M --repo-type dataset -
BODHI-S
A clinician-built graph of 779 conditions, 4,037 symptom variants and 39 specialities.
hf download ekacare/BODHI-S --repo-type dataset -
Eka MCP Server
Ground an LLM in Indian drug and treatment-protocol data over MCP.
uvx eka_mcp_server -
Parrotlet-a-en-5b
Speech recognition for English medical speech in Indian clinics.
hf download ekacare/parrotlet-a-en-5b -
Parrotlet-v-lite-4b
A 4B vision model for parsing lab reports and prescriptions from Indian medical records.
hf download ekacare/parrotlet-v-lite-4b -
Denoising Impact Evaluation
Paired noisy and denoised medical speech, from our study showing that denoising hurts ASR.
hf download ekacare/denoising-impact-evaluation-dataset --repo-type dataset -
Eka-IndicMTEB
2,532 doctor-verified medical terms in 8 Indian languages, for testing embedding models.
hf download ekacare/Eka-IndicMTEB --repo-type dataset -
Parrotlet-e
Embeddings that link medical terms across English and 12 Indian languages.
hf download ekacare/parrotlet-e -
Eka Medical ASR Evaluation
3,900+ English and Hindi recordings of medical speech, for testing speech recognition.
hf download ekacare/eka-medical-asr-evaluation-dataset --repo-type dataset -
Clinical Note Generation
Doctor-patient conversations with expert-written notes, for testing AI scribes.
hf download ekacare/clinical_note_generation_dataset --repo-type dataset -
Medical History Summarisation
Real patient histories with reference summaries and expert rubrics.
hf download ekacare/ekacare_medical_history_summarisation --repo-type dataset -
Medical Records Parsing Validation Set
288 de-identified Indian lab reports and prescriptions, annotated by experts.
hf download ekacare/medical_records_parsing_validation_set --repo-type dataset -
NidaanKosha
6.8 million lab readings from 100,000 Indian lab reports.
hf download ekacare/NidaanKosha-100k-V1.0 --repo-type dataset -
Spandan
1 million+ pulse (PPG) signals recorded on phone cameras across India.
hf download ekacare/spandan-1M-V1.0-raw --repo-type dataset
No projects match.
What's new
- POS Terminal Mock: build a card-machine integration without the machine
- Paged.js table pagination fixes
- Parrotlet-a 2.5 Pro: medical speech recognition in five Indian languages
- Open weights for medical document intelligence: Parrotlet-v 2.5 Pro, Med Doc Classifier and document-pii-redactor
- BODHI: open clinical knowledge graphs
- Pagify SDK 1.4.3 on npm
Contribute
Each project takes issues and pull requests on its own repository. Found a bug, or want to fix one? Start there.