Skip to main content

Platform Overview

Texterz is a modular AI agent infrastructure built for agencies operating multiple client deployments.

Instead of building isolated bots per channel or per client, Texterz provides a shared backend with strict tenant isolation, allowing the same system to power many white-labeled agents across channels.

This page explains the core system components and how they interact at runtime.

Core System Components​

Texterz is composed of five primary layers:

  1. Knowledge
  2. Agents
  3. Channels
  4. Tools
  5. Outbound Campaigns

Each layer has a single responsibility and can be combined without duplicating logic.

1. Knowledge (Source of Truth)​

Knowledge defines what an agent is allowed to know.

Knowledge Management

In Texterz, knowledge is stored in Knowledge Buckets, which act as controlled context containers.

  • Supported formats: PDF, DOCX, TXT, Web URLs
  • Knowledge is indexed and reused across channels
  • Buckets are scoped per client and per organization
  • Multiple agents can reference the same bucket

Knowledge never talks to users directly.
It is only accessed through an agent.

→ Go to Knowledge Guide

2. Agents (Decision & Behavior Layer)​

Agents are the runtime logic layer.

Agent List

They define:

  • Goals (e.g. qualify leads, answer support questions)
  • Behavioral constraints and tone
  • Which knowledge buckets can be accessed
  • Which tools can be triggered

Agents are channel-agnostic.
The same agent can operate on WhatsApp, Web, Email, or Social without modification.

→ Go to Agent Guide

3. Channels (Transport Layer)​

Channels connect agents to end users.

Channel Configuration

  • Channels handle message delivery and formatting
  • Business logic stays inside the agent
  • No channel-specific duplication of prompts or knowledge

Supported channels:

  • WhatsApp
  • Web Widget
  • Email
  • Instagram
  • Telegram

Agents remain unchanged regardless of where the conversation happens.

→ Go to Channels Guide

4. Tools (Execution Layer)​

Tools allow agents to perform specific tasks, not just respond with text.

Tool Configuration

Example tool outcomes:

  • Schedule appointments
  • Submit structured data
  • Trigger internal workflows

Key characteristics:

  • Tools are explicitly enabled per agent
  • Tool access is permissioned and controlled
  • Tools are executed outside the LLM

Currently supported:

  • Calendly (booking & scheduling)

→ Go to Tools Guide

5. Outbound Campaigns (Proactive Messaging)​

Outbound Campaigns enable agent-driven outreach.

Outbound Overview

  • Send messages across supported channels
  • Reuse existing agents and knowledge
  • Campaigns run within the same tenant and permission model

Typical use cases:

  • Lead follow-ups
  • Reactivation campaigns
  • Notifications and reminders

→ Go to Outbound Guide

Multi-Tenant Architecture​

Texterz is built for agencies managing multiple clients.

  • Each client operates as a Sub-Organization
  • Data, knowledge, agents, and credits are isolated
  • Branding and domains are configurable per client

This allows agencies to scale deployments without shared state or cross-client leakage.

Runtime Flow (Simplified)​

At runtime, every interaction follows the same flow:

  1. A message enters through a channel
  2. The assigned agent receives the message
  3. The agent queries allowed knowledge buckets
  4. Optional tools are triggered if required
  5. A response is returned through the same channel

This flow remains consistent across all channels.

Next Steps​

Start with the foundation:
1. Managing Knowledge