taliiq
App
Taliiq

Core Concepts

How Agents, Channels, and Tools work together in Taliiq

Taliiq separates what the agent is, where it runs, and what it can do. Your dashboard sidebar follows the same model.

The three pillars

Agents

An agent is a published configuration (active, not draft) with at least:

PiecePurpose
System promptPersonality, rules, and business logic
CapabilitiesVoice, chat, or both
BrainSystem prompt, model, first message, knowledge files
VoiceSpeech-to-text, language model routing, text-to-speech (voice agents)
First messageHow the agent greets the user

You create agents under Build → Agents. The quickstart walks through the wizard without requiring custom integrations.

Channels

A channel binds one agent to a delivery surface:

ChannelTypical use
Web widgetChat or voice on your site
SIP / phoneInbound or outbound telephony

An agent must be deployed before a channel can use it. One agent can power multiple channels.

Tools

Tools are optional. They let the agent call your APIs or MCP servers mid-conversation.

Attach tools on the agent Actions tab, or manage shared tools under Build → Tools.

Supporting pieces

AreaRole
Voice LibraryPick voices for the TTS pipeline
FilesKnowledge documents for RAG or inline context
ContactsOrg-wide contact book for widget, outbound, and CRM
IntegrationsBring your own providers — not required for first tests
LogsSessions, webhooks, and audit history

Typical lifecycle

Design — Create an agent in the wizard; test in Agent Console.

Deploy — Mark the agent active so channels and campaigns can use it.

Connect — Add any channel (web widget, SIP, phone, or outbound) pointing at that agent.

Operate — Monitor Logs; iterate on prompt, tools, or knowledge files.

Automate — Use API Keys and the API Reference for production integrations.

Member docs match the live dashboard: Build, Operate, Manage, and Observe in the left sidebar.