Skip to content

For developers

Documentation to build with ValorBrain

The shortest path to connect an agent: MCP with a bearer token, a REST API with an OpenAPI contract, and the agent-ready layer of this domain. Only resources that are live made it onto this page.

Start here

Connect your agent over MCP

Add the ValorBrain server to your MCP client configuration and call whoami on first connect to confirm the workspace. The vbm_ token is created by a human in workspace settings and carries the tenant and the permissions of its owner.

.well-known/mcp.jsonhttps://valorbrain.com.br/.well-known/mcp.json
{  "mcpServers": {    "valorbrain": {      "url": "https://mcpbrain.valor.digital/mcp",      "transport": "streamable-http",      "headers": {        "Authorization": "Bearer vbm_<seu-token>"      },      "description": "ValorBrain shared company memory: retrieve, store, handoffs, audit. Token created by a human in workspace settings (https://valorbrain.valor.digital/settings/mcp-tokens); it carries the tenant and the permissions of its owner.",      "docs": "https://valorbrain.valor.digital/auth.md"    }  }}

REST API

A REST API with an open contract

The public SaaS API answers at this base URL and the full contract is OpenAPI 3.1. The browsable reference lives inside the app, behind the workspace login.

Base URL

https://valorbrain.valor.digital

OpenAPI specification
openapi.json
Browsable reference (workspace login)
api-docs
API service catalog
api-catalog

Authentication

Bearer token in the Authorization header. The same vbm_ token works on MCP, memory REST routes and SaaS ingest. auth.md documents each case, including OAuth and a device flow for headless clients.

Read auth.md

First step

The sandbox is the product

There is no separate test environment: you try it in your own workspace, with one real project and the agents you already use. One free project to prove it in your real workflow; the paid steps have their prices published at /pricing.

The link goes through the product login and returns to the connection onboarding.

For AI agents

This site speaks to machines too

The public domain is agent-ready: pages exist as clean Markdown at the same URL with a .md suffix (full index in llms.txt), and the discovery files live at the root.