# Kaelio: governed context for data agents

Canonical: https://mudpie.ai/companies/kaelio/
Breadcrumb: [Home](https://mudpie.ai/) / [Companies](https://mudpie.ai/companies/) / [Kaelio: governed context for data agents](https://mudpie.ai/companies/kaelio/)
Author: Ali Abouelatta (https://mudpie.ai/authors/ali-abouelatta/)
Published: 2026-09-19
Updated: 2026-09-19
Research type: Company profile
Method: Company and accelerator sources checked 2026-09-20. Product claims are attributed to their sources; this is research, not a hands-on product trial.

## What it does

Kaelio builds ktx, an open-source context layer for data agents. It ingests warehouse metadata, BI definitions, modeling code, query history and business documentation, then turns that evidence into reviewable Markdown knowledge and executable YAML metric definitions. Agents can ask ktx for a metric, let it resolve grain and joins, and receive compiled read-only SQL instead of guessing from table names. The [YC profile](https://www.ycombinator.com/companies/kaelio) and [homepage](https://www.kaelio.com/) describe the problem as reliable data-agent context.

The fit is an analytics team that has already seen plausible SQL produce the wrong number. Kaelio is not another dashboard or a replacement for dbt. It is a context and semantic layer intended to make agent queries governed, explainable and repeatable.

## Why I’d look closer

The technical boundary is unusually legible. ktx keeps definitions and caveats in files that can be reviewed and merged through git, while the query engine handles joins, grain, filters and read-only execution. The homepage names Snowflake, BigQuery, Looker, dbt, Notion, MCP agents and common data tools. The [pricing page](https://www.kaelio.com/pricing) lists the open-source engine as free; hosted Cloud and managed Data Agent are contact-led.

The founders have direct data and AI-safety context. The [YC biographies](https://www.ycombinator.com/companies/kaelio) describe Luca Martial with enterprise data systems and AI-safety experience and Andrey Avtomonov as a repeat AI-safety founder with CERN and Dataiku engineering backgrounds.

## What I’d ask

How does ktx detect stale definitions, conflicting metric owners and dangerous joins, and who approves a proposed context update? I’d connect a read-only warehouse, replay known analytics questions and compare agent-generated SQL to the team’s canonical answers before exposing write or workflow actions.

## My editorial take

Kaelio is a strong fit for organizations that want data agents to be reliable enough for real decisions. The open-source files and review loop are the important parts. The product earns trust when the context layer makes wrong answers easier to catch than to explain away.

## Quick facts

| Field | Sourced detail |
|---|---|
| Buyer fit | Analytics/data teams using AI agents over warehouses and BI tools |
| Open-source plan | ktx engine free under Apache 2.0; self-hosted |
| Hosted options | ktx Cloud and managed Data Agent; contact-led |
| Runtime | CLI/MCP, read-only SQL, reviewable Markdown/YAML context |

## Sources checked

Checked 2026-09-20.

| Source | Used for |
|---|---|
| [YC company profile](https://www.ycombinator.com/companies/kaelio) | Product, founders and ktx launch description |
| [Kaelio homepage](https://www.kaelio.com/) | Current context/query workflow and integrations |
| [Kaelio pricing](https://www.kaelio.com/pricing) | Open-source/Cloud/managed plan structure |
| [ktx docs](https://docs.kaelio.com/ktx/docs/getting-started/introduction) | Context-as-code, CLI/MCP and review model |

## Cohort context

Kaelio is listed in Spring 2025. In our 2026-09-18 directory snapshot, 97 of 143 listed companies in that cohort have YC’s primary industry label B2B (67.8%). This is a current-directory comparison, not an original intake count or a performance ranking. [Nine-cohort dataset](https://mudpie.ai/research/yc-cohorts-2026-09-19.json).

## Public website snapshot

Observed 2026-09-19T16:15:41.438Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.

| Signal | Homepage observation |
| --- | --- |
| Product description metadata | Observed |
| Canonical link | Observed |
| H1 or H2 heading | Observed |
| Typed structured data | Observed |
| Docs/developer link | Observed |
| Pricing link | Observed |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |

[Public observations](https://mudpie.ai/research/yc-homepage-links-2026-09-19.json) · [Collection method](https://mudpie.ai/research/yc-homepage-methods/README.md). Missing links here do not establish that a capability or file is absent elsewhere.


## Author disclosure

I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.