Integration Guides
rFabric is designed for coexistence. Mature robotics teams already have tools for logging, visualization, training, storage, and operations. The platform is built to become the lifecycle system of record across those tools, not to demand a one-week full-stack replacement.
Integration Domains
Robot data and middleware
- ROS and ROS 2
- rosbag and rosbag2
- MCAP
- custom robot-side data recorders
- simulator-generated sessions when promoted into managed assets
Storage and data infrastructure
- S3, GCS, Azure Blob
- local batch export and import paths
- edge cache or relay systems
- structured manifest exchange with external training or analytics systems
Visualization and review tools
The platform is designed to preserve lineage and review continuity even when specialist tools remain in use.
- Foxglove
- Rerun
- internal viewers where teams still rely on them
Training and MLOps systems
- internal PyTorch training repos
- notebook-driven workflows
- cloud and on-prem schedulers
- external experiment dashboards where teams are not ready to migrate fully
Operations systems
- chatops and on-call routing
- issue tracking
- maintenance and service systems
- customer-specific operational tooling
How Integration Works
Start with the painful boundary
The best first integration is usually the most fragmented one:
- upload and review
- curation and dataset finalization
- release evidence and approval
- rollout and fleet visibility
Preserve one identity model
Even when external tools remain in use, keep entity IDs, the lineage chain, and workflow history anchored in the platform so context does not fragment.
Replace glue before replacing specialists
The highest-value early move is often to replace brittle handoffs and spreadsheet-driven coordination, not every specialist interface.
Expand as the system of record proves itself
As teams trust the platform’s lineage, approvals, and operational visibility, more lifecycle stages can move under the same control plane.
Why This Matters
Incremental adoption is an architectural requirement, not an afterthought. A robotics platform that cannot coexist with the existing stack will struggle to become the source of truth for the lifecycle. A platform that can coexist, while preserving entity continuity and governance, can earn a deeper role one lifecycle stage at a time.