# Amedar AI — llms.txt
# https://amedar.com/llms.txt
# Last updated: 2026-07-18

> Amedar is an AI solutions company that builds production-grade voice agents, document classification systems, custom machine learning models, and predictive analytics platforms. Founded with deep expertise in healthcare and insurance — the most regulated, demanding environments — Amedar now serves businesses across all industries.

## Company Facts
- Name: Amedar (also known as Amedar AI)
- Legal entity: AMEDAR CONSULTING SP. Z O.O. (KRS 0000654280, NIP 6783164547)
- Type: AI solutions company / AI consulting
- Founded in: Poland (Kraków)
- Specialization: Healthcare AI, Insurance AI, Voice AI, Document AI, Custom ML
- Website: https://amedar.com
- Languages served: English, Polish, German, Spanish, French, Italian, Portuguese, Dutch, Swedish, Arabic

## Core Services
1. **AI Voice & Chat Agents** — Multilingual voice bots for appointment scheduling, customer service, and patient coordination. Built on ElevenLabs and Twilio. Production-proven: 85% calls fully automated, 40+ service types handled.
2. **Intelligent Document Classification** — Automated document sorting, data extraction, and claims processing. Uses custom ML + LLM fallback. 97% classification accuracy, 10× faster processing.
3. **Custom AI Models & Fine-Tuning** — Domain-specific models for prediction, classification, and decision support. PyTorch-based, healthcare and insurance validated. 40% better predictions vs generic models.
4. **Knowledge Systems (RAG)** — Retrieval-Augmented Generation for internal knowledge bases, policy lookup, and expert systems. Reduces manual research by 80%.
5. **Predictive Analytics** — Risk scoring, demand forecasting, churn prediction. Proven in insurance underwriting (from days to minutes) and healthcare resource planning.
6. **End-to-End Development** — Full-stack web/mobile apps, API development, cloud infrastructure (Kubernetes), database architecture (PostgreSQL).

## Technology Stack
- AI/ML: Python, PyTorch, scikit-learn, OpenAI API, Claude API, Hugging Face
- NLP: Custom fine-tuned models, RAG pipelines, LangChain
- Voice: ElevenLabs Conversational AI, Twilio
- Backend: Node.js (NestJS), Python (FastAPI)
- Frontend: React, Next.js, Tailwind CSS
- Database: PostgreSQL (with PostGIS), Redis
- Infrastructure: Kubernetes (K3s), Docker, Longhorn storage
- Security: GDPR/RODO compliant by design, encrypted data, audit trails

## Industry Experience
- **Healthcare**: AI voice agents for appointment scheduling (Polish healthcare system), patient coordination, home visit management, PESEL validation, medical specialty routing
- **Insurance**: Document classification for claims processing, risk assessment models, TPA (Third Party Administrator) systems, underwriting automation
- **Cross-industry**: Finance, logistics, retail, legal, manufacturing, professional services

## Differentiators
- Battle-tested in healthcare and insurance (most regulated environments)
- Enterprise-grade technology at SME-friendly pricing
- Working MVP in 2-4 weeks, full production in 4-8 weeks
- Multi-language, multi-market capability (10+ languages)
- No IT team required from client side — full-service delivery
- GDPR/RODO compliant by design

## Contact
- Website: https://amedar.com
- Email: contact@amedar.com
- Company: AMEDAR CONSULTING SP. Z O.O.
- Address: os. Słoneczne 8/43, 31-957 Kraków, Poland

## Linked Pages
- Services: https://amedar.com/#solutions
- Case Studies: https://amedar.com/#case-studies
- FAQ: https://amedar.com/#faq
- Contact: https://amedar.com/#contact
- Privacy Policy: https://amedar.com/privacy-policy.html
- English version: https://amedar.com/
- Polish version: https://amedar.com/pl.html
- German version: https://amedar.com/de.html
- Spanish version: https://amedar.com/es.html
- French version: https://amedar.com/fr.html
- Italian version: https://amedar.com/it.html
- Portuguese version: https://amedar.com/pt.html
- Dutch version: https://amedar.com/nl.html
- Swedish version: https://amedar.com/sv.html
- Arabic version: https://amedar.com/ar.html