Vanta CLI is one REPL with two interchangeable engines underneath: Claude through the Claude Agent SDK, running in-process, or OpenAI Codex, driven as a subprocess. Both speak the same MCP tools, stream the same tool cards, and report the same way — so which engine is underneath is a detail you can change at any time.
The anthropic engine loads the Claude Agent SDK in-process: the full Claude Code toolset — Read, Write, Edit, Bash, Glob, Grep, plus skills, subagents, and hooks. The openai engine shells out to the user-installed Codex CLI — one codex exec per turn, with its own apply_patch and sandbox. Both surface identically.
Change engine, model, effort, or auth without restarting the process. /model and /effort carry the same conversation forward; /provider and /auth start a clean thread on the new engine — and the REPL tells you which.
Continuity travels as far as the provider allows: a same-engine switch resumes the session/thread id; a cross-engine switch starts fresh, and the REPL says so.
A first-party agent CLI behind one REPL and contract. Swap Claude for Codex mid-session; Vanta drives it exactly like it drives everything else.
ExplorePoint the openai engine at any OpenAI-compatible host. A built-in Chat→Responses bridge means OpenRouter, Groq, Together, and local Ollama all just work.
ExploreOne LLM plans the work, spawns agents, reviews their diffs, and merges. You talk to it like a tech lead.
ExploreVanta Studio is launching soon. Join the waitlist and start directing a team.