Column-Level Lineage
Track which source columns flow into each metric and dimension — column-level, not just table-level. When a source column changes, you know exactly which downstream definitions are affected before running anything.
Platform
One platform. Semantic layer, ETL, and agentic routing.
Loomkindle unifies semantic model authoring, agentic ETL, and column-level lineage tracking in a single YAML-first platform. It works alongside dbt, Airflow, Dagster, and your cloud warehouse — not instead of them. The agent handles the parts that are currently falling through the cracks: schema drift, metric consistency, and routing decisions that nobody wants to debug at 3am.
YAML-first
Everything in git
16+
Native integrations
Auto
Schema drift handling
Semantic Layer
The Semantic Layer
Define your metrics, dimensions, and entity relationships once in YAML. Loomkindle's semantic layer exposes these definitions consistently to every downstream consumer — dbt models, BI tools like Sigma and Hex, and LLM-based data agents. When a definition changes, update one YAML file, open a pull request, and every downstream tool gets the updated definition on next sync. No more Looker LookML and dbt Semantic Layer drifting apart.
- Metric definitions that survive schema changes
- Consistent entity resolution across tools
- Version-controlled in git with full diff history
- Agent-aware context for LLM-based queries
Agentic ETL Engine
Agentic ETL Engine
Loomkindle's ETL engine isn't rule-based — it's agent-based. When your upstream schema changes, the agent detects the drift, evaluates impact on downstream models, and reroutes the transform DAG automatically. No on-call page. No manual intervention.
- Source connectors for major cloud warehouses
- Schema drift detection and automated remediation
- Dependency-aware transform scheduling
- Every routing decision logged and auditable
Agentic Routing
How Agentic Routing Works
The routing agent runs continuously, watching for changes and making decisions — not just when you tell it to.
Monitor
Agent continuously watches upstream schemas for column additions, renames, type changes, and deletions.
Evaluate
Impact analysis identifies which downstream transforms and semantic definitions are affected by the change.
Reroute
Affected transforms are automatically rerouted using the updated schema. The semantic layer remains consistent.
Log
Every routing decision is logged with full context — what changed, why the reroute was made, which models were affected.
Observability & Lineage
Observability & Lineage
Routing Decision Log
Every agent decision is logged with timestamp, trigger, affected models, and the action taken. Fully auditable.
Data Quality Integration
Designed to work alongside Monte Carlo and Great Expectations. Loomkindle handles routing and semantic consistency; your observability layer handles anomaly detection and data quality SLAs. They're complementary, not competing.
Ready to see the platform?
Request early access or read the quickstart guide to get started.