Getting Started

Introduction

What Vettero is and how you use it in the web app.

Vettero is a web app for building and running AI experiences in a workspace. You design agents and workflows, connect them to chat and voice channels, ground them in your data, and operate everything from one place — without writing application code.

Open Vettero in the browser, sign in, and work inside a workspace. Everything you create — agents, workflows, knowledge bases, tables, channels, and integrations — lives there and stays isolated from other workspaces.

Invite people at company Users, then add them under Configuration → Workspace Settings → Users. Assign roles so they can use tables, in-app chat, widgets, and approvals. See Workspace users.

What you can do

Agents

Conversational or general AI with tools, knowledge, prompts, guardrails, and channel-aware behavior.

Workflows

Visual graphs of steps — conditions, loops, data I/O, messaging, and human-in-the-loop approvals.

Studio

In-app AI assistant that helps you create and edit agents, workflows, and other workspace resources by chat.

Channels

Reach people on in-app chat, WhatsApp, email, voice, and more.

Knowledge bases

Index documents and data so agents can retrieve and answer from your content.

Tables

Structured workspace data that agents, queries, and workflows can read and write.

Forms

Table layouts and private links for collecting or updating rows without the grid.

Queries

Saved table operations agents, functions, widgets, and workflows can call.

Functions

Sandboxed JavaScript or Python with schemas — callable from agents, workflows, and widgets.

Skills

Markdown playbooks agents load on demand with get_skill.

Widgets

Hosted or URL iframe pages on the home dashboard and sidebar.

Chat widgets

Liquid HTML UIs inside an agent chat message, with click events back to the AI.

Prompts

Tips for writing agent prompts, choosing an LLM, and tuning other agent settings.

Variables

Liquid templates and path bindings for dynamic values across prompts, workflows, and configs.

Environment variables

Workspace key/value config and secrets referenced as env.*.

Approval flow

Pause tools and workflow steps until someone approves, rejects, or submits information.

Notifications

In-app bell inbox for members — from agents, workflows, or the Send Notification block.

Companion app

iPhone and Android app for messages, approvals, alerts, tables, and in-app agents.

Files

Upload and manage workspace files for agents, tables, chat, and workflows.

Email templates

Reusable Liquid email bodies in MJML, HTML, or Markdown.

Text templates

Reusable Liquid text rendered to a string, with HTML or Markdown preview.

WhatsApp templates

Twilio WhatsApp content templates for outbound and in-session messages.

APIs

Import OpenAPI schemas so agents and MCP can call REST operations.

Chat integrations

Web widget and in-app sidebar chat for agents and chat workflows.

MCP integrations

Connect outbound to external MCP servers and use their tools in agents.

MCP servers

Host a workspace MCP endpoint that exposes selected resources as tools.

Chat profiles

Human-agent profiles for handover, members, quick replies, and transfers.

Workspace users

Add company users to the workspace, set Owner or Developer, and assign roles.

Roles

Nested workspace roles for access control and approval gating.

Trash

Restore or permanently delete soft-deleted workspace resources.

Schedule

One-time runs of agents or workflows at a datetime or after a timeout.

Event handlers

Recurring or trigger-based runs of agents and workflows.

HTTP servers

Inbound HTTP APIs grouped under a base URL slug.

Third party services

Connect LLM providers and external APIs used by channels, agents, and workflows.

Core concepts

Concept	Meaning
Workspace	Your project boundary — agents, workflows, data, channels, members, and settings live here.
Agent	An LLM-driven assistant with a task prompt, tools, optional knowledge, and a channel (chat, voice, or general).
Workflow	A visual graph of blocks that automates multi-step processes, often tied to a channel.
Channel	How people reach an agent (in-app chat, WhatsApp, email, voice, and similar).
Connected service	Credentials for models and external APIs used by agents and workflow steps.
Studio	Chat-based builder inside Vettero for creating and changing workspace resources.
Workspace user	A company user added to this workspace, with Owner/Developer flags and optional roles.

How it fits together

Build — Create agents, workflows, functions, skills, knowledge bases, and tables in the web app (or ask Studio to help).
Connect — Attach LLM providers and third-party services, then set up channels so users can message or call in.
Run — People talk to your agents on those channels; workflows and tools act on your tables, queries, and APIs.
Operate — Preview in the builder, review conversations, use approvals and notifications when a human must step in, and manage workspace users over time. The companion app is the mobile inbox and workspace hub for that operate step.

Where to go next

Sign in, open a workspace, connect an LLM provider, and add workspace users if you work as a team.
Create an agent and preview it in the builder.
Put it on in-app or web chat, or on WhatsApp, email, or voice.
Walk through a full example on Use cases.
Install the companion app so the team can handle chats, approvals, and alerts on a phone.

Browse Build, Data, Integrations, and the other sections for each building block as you expand the workspace.

Use Cases

Multi-step examples of what you can build with Vettero and how the pieces fit together.