Emcie
Emcie provides an optimized inference platform and open-source conversational AI engine (Parlant) for building reliable, production-grade AI agents with structured workflows and oversight tools, designed for customer-facing applications in high-stakes industries.
At a Glance
- Generative AI development teams
- Software engineers building conversational AI
- Financial services companies
- Healthcare organizations
- +6 more
AI Tools by Emcie
(2)
Emcie
Optimized Inference for Parlant Agents
Parlant
LLM Agent Alignment Engine
Discussions
No discussions yet
Be the first to start a discussion about Emcie
Latest News
Emcie launches auto-optimizing inference platform for Parlant agents, cutting costs by up to 10x
Parlant 3.1 released with OpenTelemetry integration, guideline criticality levels, and enhanced control mechanisms
Parlant and Pipecat announce official voice agent integration
Parlant surpasses 12,000 GitHub stars, demonstrating strong open-source community adoption
Products & Services
Open-source conversational AI engine (Apache 2.0 license) for building reliable, production-grade chat agents. Replaces brittle prompt engineering and flow-based logic with behavioral guidelines that trigger actions like tool calling or human routing. Designed for regulated or sensitive sectors including financial services, healthcare, legal, and government.
Auto-optimizing inference platform for Parlant agents that reduces operational costs through dynamic prompt tuning and Small Language Model (SLM) distillation. Maintains response accuracy by optimizing individual NLP tasks independently. Achieves LLM-level accuracy at 5 to 10 times lower operational costs through a Teacher/Student model approach.
Market Position
Positions as a middle layer between traditional rigid chatbot flows and unpredictable pure LLM agents. Differentiates through emphasis on control, reliability, and behavioral modeling rather than just prompt engineering. Targets production-grade, customer-facing agents in high-stakes industries where consistency and compliance are critical. Open-source approach with Apache 2.0 license contrasts with proprietary and open-core competitors.
Leadership
Founders
Yam Marcovitz
CEO & Co-Founder. Previously Senior Architect at Microsoft (2021-2022), Senior Staff Software Engineer at Apex.AI (2022-2023), Senior Software Engineer at Microsoft (2021), System Architect at Dataloop AI (2017), System Architect at Lucidlogix Technologies (2016-2017), and Researcher at Lucidlogix (2014-2016). Experience in mission-critical systems and real-time software engineering.
Dor Zohar
CTO & Co-Founder. Previously Senior Software Engineer at EverC (2022-2024), Engineering Technical Lead at Furious Corp. (2018-2022), Fullstack Developer at Cronus Cyber Technologies (2016-2017). Background in mission-critical systems outside of traditional NLP.
Executive Team
Yam Marcovitz
CEO & Co-Founder / Lead Architect @ Parlant.io
Software engineering background with experience at Microsoft, Apex.AI, and expertise in mission-critical systems
Dor Zohar
CTO & Co-Founder / Lead Developer @ Parlant.io
Engineering and technical leadership background with experience at EverC and Furious Corp.
Board of Directors
Founding Story
Founded in early 2023 by Yam Marcovitz and Dor Zohar when GPT-4 was released. The founders, coming from backgrounds in mission-critical and real-time systems at companies like Microsoft and Check Point, were excited about building customer-facing AI agents but frustrated with existing solutions. They found traditional chatbots too rigid and expensive, while purely LLM-driven agents lacked reliability and alignment with business requirements. Their mission-critical systems experience made them value ultimate control and determinism, which they found lacking in NLP. They set out to bring that same rigor to conversational AI by creating a hybrid approach combining the flexibility of LLMs with structured controls.
Business Model
Revenue Model
Dual model: (1) Open-source Parlant framework - fully free under Apache 2.0 license with no plans to monetize the core product or implement freemium model. (2) Commercial Emcie inference platform - usage-based API pricing for optimized inference services with cost reduction from Teacher to Student model pricing.
Pricing Tiers
Starting credit for new users of Emcie platform
General-purpose model tier optimized for cost efficiency. Approximately $0.02-$0.05 per message.
Heavy-duty model tier for more complex use cases requiring higher accuracy.
Open-source under Apache 2.0 license. Fully free to use, modify, and deploy including commercial projects.
Target Markets
- Generative AI development teams
- Software engineers building conversational AI
- Financial services companies
- Healthcare organizations
- Legal firms
- Government agencies
- Customer service chatbots for high-stakes industries
- Financial advisory assistants
- Healthcare agents for patient engagement
- Legal sector conversational AI
- Government agency support systems
- SaaS product technical support