A self-evolving ontology layer for data agents integrating via MCP tools.
Bridge the agent-data gap across heterogeneous tables, files, and databases with MCP integration.
EvoOntology is an open-source ontology layer designed to bridge the gap between data agents and heterogeneous data sources. It provides active semantic access through Model Context Protocol (MCP) tools and continuously evolves its schema based on execution trajectories. It suits data science professionals and engineers building reliable data agents on Claude Code or Codex.
It acts as a trainable agent state—not model weights—comprising Content, Schema, and Tool layers. The builder initializes semantic objects from your workload, while an evolution agent proposes bounded updates and validates every candidate against its parent using gated versioning.
It suits data scientists and machine learning engineers deploying autonomous agents on complex databases and files who struggle with semantic uncertainty. It does not suit teams unwilling to manage iterative agent evaluation loops and workload-grounded validation pipelines.
Integrates seamlessly with Claude Code and Codex via Model Context Protocol (MCP) tools.
Builds a tailored semantic layer directly from your raw data and workload evidence.
Refines the ontology layer automatically based on interaction history and trajectory diagnostics.
Publishes candidate ontology patches only when paired evaluation shows reproducible improvements.
Free under its open source license, with no paid hosted option mentioned in the repository. The only cost is running the tooling and evaluation workflows yourself.
883 open remote AI-training roles, updated November 16, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
These reviewed AI tools do the same job instantly — from our scored directory.
See Zekai first in Google