Marker provides a platform for observing real voice and chat agents, running lifelike simulations, and evaluating every interaction with shared Markers. It supports on-prem and air-gapped deployments to keep data in customer environments. The system includes loops for improving both agents and the evaluation process itself by routing machine judgments to humans for labeling, measuring agreement, and refining monitors and simulations. It is designed for teams building production voice agents who need reliable, grounded evaluation in their own cloud or on-premise infrastructure.
Marker is a LLM eval & observability project. It focuses on evaluating and improving the quality of voice and chat agents without reliable automated judges or consistent human feedback. Marker is a B2B product aimed at AI agent developers and ML teams. Marker is sold on an enterprise-only basis. It runs on the web and API, and it can be self-hosted.
It is developed by Marker (United States), and it first shipped in 2025. Among its 6 catalogued features are on-prem deployment, agent simulation, and human-in-the-loop labeling. It exposes integrations via a public API.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Same category — not a similarity match