Savalera Agents Hub: Internal Deployment of X-Series Agents Now Live
We've deployed our first agent platform featuring ax1 and vx1 architectures for internal use across research, data analysis, and content creation workflows.
We’ve launched the Savalera Agents Hub, our internal platform for deploying and testing X-Series agents in real-world scenarios. The hub currently hosts ax1 and vx1 agent configurations and serves as both a development testbed and practical tool for our daily operations.
Current deployment
The platform is operational with both fast (ax1) and deep (vx1) agent architectures handling various internal tasks including research scenario modeling, RAG (Retrieval-Augmented Generation) implementations, data analysis workflows, and content creation projects.
We’re running agents on both open-weight and commercial LLM backends, allowing us to test architecture performance across different model capabilities and cost profiles. This multi-backend approach helps us understand how our project management-based configurations perform with various underlying language models.
Internal testing approach
The hub serves as our primary testing environment where commercial agent development happens alongside actual usage. This approach ensures our agents face real operational challenges rather than just synthetic benchmarks, providing authentic feedback for architecture refinement.
Our team uses the platform daily for research tasks, documentation, and analysis work, generating practical insights into agent reliability, workflow efficiency, and user experience patterns that will inform commercial development.
Technical infrastructure
The platform features a deployed user interface designed for internal team access and experimentation. The architecture supports rapid agent configuration changes and testing of different workflow approaches as we refine our project management methodologies.
The hub infrastructure is designed to scale toward external access, though current deployment focuses entirely on internal validation and development needs.
Development insights
Using our own agents for daily work provides direct feedback on performance characteristics, failure modes, and workflow integration challenges. This internal-first approach helps identify gaps between theoretical capabilities and practical utility.
The multi-agent architecture allows us to compare fast execution (ax1) versus deeper planning approaches (vx1) across different task types, informing decisions about which architectures to deploy for specific use cases.
Future plans
The Agents Hub will eventually serve as a demonstration platform for customers to experience our agent capabilities. However, current focus remains on internal development and validation before considering external access.
This deployment represents the first practical implementation of our X-Series architectures in a production-like environment, moving beyond development testing to actual operational use.
The Savalera Agents Hub supports our agent development process by providing real-world testing environments for commercial product validation.