AI Orchestration

Design, configure, and observe AI agents — from building custom agents to monitoring every token they consume.

Overview

AI Orchestration

Agents

Architect

Model Center

Prompt Lab

Knowledge

Observe

Available Agent Types

Standard

Single-step LLM reasoning

Reasoning

Chain-of-thought pipeline

Autonomous

Tool-calling multi-step

Observability: 4 runs today · 2,418 tokens · 0 errors

The AI Orchestration mode is TotalApp's platform for building, managing, and observing AI agents at scale. Where the Automation mode provides the workflow canvas, AI Orchestration provides the infrastructure underneath: which models are available, how agents behave, what tools they can call, what knowledge they have access to, and what they actually did.

Switch to AI Orchestration mode from the mode switcher in the top header. The sidebar updates to show all orchestration screens.

Start with the Agents Screen

Open the Agents screen to see all pre-built agent types available in TotalApp. Drag any agent node from there into the Workflow Editor to instantly add AI reasoning to a pipeline — no configuration required for the defaults.

AI Orchestration Screens

Agents

Browse and configure pre-built agent nodes — Standard, Reasoning, Autonomous, Multimodal, Domain-Specific, and more.

View documentation →

Agent Architect

Design your own AI agents — choose LLM, define prompt chains, and wire them into any workflow node.

View documentation →

Model Center

Manage which local LLMs are active — TinyLlama, Claude, and more — and switch between models from one place.

View documentation →

Prompt Lab

Test and refine prompts in a live sandbox before wiring them into your workflows.

View documentation →

Knowledge Base

Upload PDFs and text documents, convert them to vector embeddings, and make them available to any AI agent.

View documentation →

Agent Persona Editor

Define each agent's character, capabilities, and constraints with detailed system prompts.

View documentation →

Tool / Function Registry

Define JSON Schema tools — database queries, API calls, file I/O — and wire them into any agent in seconds.

View documentation →

Observability & Logs

See what every AI agent did: tokens used, tools called, errors thrown — across all apps and users.

View documentation →

How AI Orchestration Relates to the Workflow Editor

The Workflow Editor (in Automation mode) is where you assemble pipelines using agent nodes. AI Orchestration is where you configure what those agent nodes do:

Agent Architect defines the agent's LLM, system prompt, and tool list — these settings travel with the agent node when you drag it into a workflow.
Model Center controls which models are available to select in Agent Architect and Prompt Lab.
Knowledge Base provides the vector store that RAG-enabled agents query at runtime.
Tool Registry exposes callable functions that agents can invoke during a workflow run.
Observability shows the execution trace of every agent call — tokens, latency, tool calls — after the workflow runs.

Mode Relationship

You do not need to switch modes to use agents in a workflow. Agents configured in AI Orchestration mode are automatically available as node types in the Workflow Editor, regardless of which mode you are currently in.