ktx vs WrenAI
Self-improving context layer for data agents — ingests dbt/Looker/wikis, maps your warehouse, builds a semantic layer with approved metrics, and serves Claude Code/Codex via CLI and MCP. — versus — Open-source GenBI engine: agents write governed SQL and deploy shareable dashboards over 22+ data sources, grounded in a Git-friendly context layer (MDL semantics, definitions, memory).
Both are governed semantic layers that make agents trustworthy over company data: ktx curates warehouse metrics and serves read-only SQL context; Wren goes further into governed execution and agent-deployed dashboards.
| ktx | WrenAI | |
|---|---|---|
| Stars | 1.5k | 16k |
| Forks | 93 | 1.8k |
| Language | TypeScript | Python |
| License | Apache-2.0 | NOASSERTION |
| Last activity | 7 days ago | 2 days ago |
| Topics | rag, agents | agents |
| Curated connections | 3 | 2 |
ktx — the curator's take
Reach for it when agents re-explore your warehouse on every question and invent their own metric logic: ktx samples tables, detects joinable columns (resolving chasm/fan traps), absorbs dbt/MetricFlow/LookML/Notion knowledge into one searchable surface, and flags contradictions for human review. Read-only by design; runs locally on your own LLM keys or your Claude Code / Codex login. Skip it if you have no SQL warehouse to sit on, or for one ad-hoc query. It ingests your existing semantic layers rather than replacing them. YC-backed (Kaelio); telemetry is on by default with opt-out.
WrenAI — the curator's take
The strongest open answer to 'my agent writes confidently wrong SQL': business definitions, approved joins and past queries live in reviewable files, not prompts, and dry-plan validation catches errors before execution. Agent-driven by design (skills + CLI, works through Claude Code/Cursor). Skip for one-off charts from a CSV — the context layer is the point, and it takes real setup.