AI Security & model governance

Last updated: September 2026

AtlasGTM is the governed operating, orchestration, context, and analytics layer around AI systems such as Claude, Gemini, OpenAI, Cursor, Codex, and MCP-connected clients. Provider training policies are stated per vendor after review — never as a blanket claim.

Anthropic — Claude

Optional model provider for Atlas Intelligence and MCP-connected workflows.

OpenAI — OpenAI API

Optional model provider.

Google — Vertex / Gemini

Optional model provider.

Ollama — Local / self-hosted models

Optional local model runtime where configured.

MCP clients — Cursor, Claude, Codex, and other MCP clients

Customer-operated agents that connect to AtlasGTM through the governed MCP gateway.

Security questions: security@atlasgtm.io