Keboola is a data platform that positions itself as an AI-ready data foundation for finance and other data-heavy teams. It covers ingestion from a stated 1,500+ data sources, SQL, Python and dbt transformations, orchestration and scheduling, governance and a data catalogue, and delivery to tools such as Tableau, Looker, Excel and Streamlit. It runs on AWS, Azure, GCP or a private cloud and integrates with Snowflake, BigQuery and DuckDB.
The AI layer has several named parts. Kai is an assistant for natural-language questions, labelled in the interface as an AI that can make mistakes. The finance pages describe Audit and Controlling Agents that watch financials for budget drift, duplicate postings and missed eliminations before the close. AI-assisted development helps write SQL, Python and dbt code. The Keboola MCP Server, released under the MIT licence and actively maintained on GitHub at the time of review, lets AI clients such as Claude and ChatGPT, and IDEs such as Cursor and Windsurf, control pipelines through the Model Context Protocol.
Pricing scales with data volume rather than seats. The Free plan costs $0 and includes unlimited pipelines, SQL and Python transformations, an extra-small Snowflake backend, the Flow Builder, the MCP Server and one project, with 120 free compute minutes in the first month and 60 a month after that; extra minutes cost $0.14 each. The Enterprise plan is quote-only and adds data sharing and a catalogue, Git-based CI/CD with SOX compliance, SAML and SSO, any storage backend and a dedicated technical account manager. The site states SOC 2 Type II, GDPR and HIPAA compliance, names customers including Home Credit, Groupon and Apify, and cites a 4.8 G2 rating from 100-plus reviews; all of these are vendor-reported.
