Building practical tools for real-world LLM systems.
We focus on the core challenges of production LLM systems:
- Reliability — detecting weak reasoning and hallucinations
- Orchestration — managing multi-step agent workflows
- RAG tooling — building and testing retrieval pipelines
- Cost & control — managing usage, access, and security
Stateful multi-agent runtime for Spring AI.
Run long-lived workflows with retries, recovery, and graph execution.
→ https://github.com/datallmhub/spring-agent-flow
CLI tool to manage and optimize RAG pipelines.
Test, debug, and iterate on retrieval workflows directly from the terminal.
→ https://github.com/datallmhub/ragctl
Control layer for LLM systems.
Manage cost, access, and security for production AI usage.
→ https://github.com/datallmhub/TensorWall
- AI reliability
- Agent systems
- Retrieval-Augmented Generation (RAG)
- Production-grade LLM tooling
Simple tools, clear responsibilities, production-first design.
Early stage — feedback and contributions welcome.