Explore APIs and MCP

The same governed functions - MCP or HTTPS

Connect approved AI environments through MCP, or invoke Maestro from internal applications over HTTPS. One specialist platform: permissions, approved data sources, defined methodologies and auditability on every call. Start in a sandbox with demonstration data, then request production when your integration is verified.

Request sandbox access Request developer access See an institutional workflow

Two surfaces, one control plane

MCP for AI environments

Connect Claude, Codex, Gemini, ChatGPT, or Copilot so teams reach specialist FX and rates capabilities without leaving the assistants they already use.

HTTPS for enterprise apps

Embed the same governed functions in internal applications, automated processes and proprietary workflows.

Auditable responses

Every result is designed to carry sources, timestamps, methodology and warnings - not just a bare number.

Simple request shape

Illustrative - exact routes and schemas are shared after access is granted.

POST /v1/workflows/hedge-compare
Authorization: Bearer <sandbox_token>

{
  "book_id": "demo-usd-mxn",
  "budget_rate": 18.40,
  "alternatives": ["do_nothing", "forward_strip", "tarf"]
}

→ decision pack
  results, data_class, sources, timestamp,
  tools_called, methodology, warnings, result_id

Onboarding path

Understand MCP and HTTPS surfaces (this page).
Review an institutional workflow example.
Create or sign in to a ZeroVaR account.
Obtain sandbox access and credentials.
Choose MCP connector or HTTPS.
Complete a verified first call.
Request production access when ready.

Request developer access

Prefer the primary path? Request sandbox access. Or tell us what you want to build - we follow up with sandbox onboarding without a password-gated portal from this form.