AI Orchestration

Build, run, and observe
AI agents at scale.

Design multi-agent pipelines visually, manage every LLM from one place, test prompts live, and monitor every invocation — all inside TotalApp.

Agent Architect Prompt Lab Model Center Observability Tool Registry Knowledge Base
Agent Architect

Design AI agents
without writing code

Pick an LLM, define prompt chains, wire in tools, and connect agents to any workflow — all in a visual canvas. Build reasoning, autonomous, and multi-modal agents in minutes.

  • Pick from local or cloud LLMs per agent — mix models in one pipeline
  • Build multi-step prompt pipelines visually with drag-and-drop nodes
  • Export agents as reusable nodes in the Workflow Editor
  • Standard, Reasoning, Autonomous, Multimodal, and Domain-specific types
Open Agent Architect
totalapp.app / orch-agent-architect
Agent Types
🤖Standard
🧠Reasoning
Autonomous
🖼️Multimodal
Tools
🔧web_search
📊query_db
📧send_email
Input
User Query
LLM
Claude Sonnet
Reasoning agent
Tool
web_search
Live retrieval
Output
Response
4 nodes · 2 tools · Claude Sonnet
Prompt Lab

Test prompts live before
wiring them into agents

Write, run, and compare prompts across any configured model in a live sandbox. Save successful templates for reuse as agent nodes — no copy-pasting between tools.

  • Run prompts against any active model — local or cloud — in one click
  • Compare outputs side-by-side across models with diff highlighting
  • Save prompt templates and inject them directly into Agent Architect nodes
  • Token count, latency, and cost shown per run
Open Prompt Lab
totalapp.app / orch-prompt-lab
Claude Sonnet 4.6
TinyLlama 1.1B
Run All
Summarize the following text in 3 bullet points. Focus on actionable insights only:
{{input}}
Claude Sonnet · 312ms · 847 tok
• Conversion rates dropped 12% in mobile checkout — optimise payment UX
• Desktop AOV is 2.3x mobile — target upsells on desktop
• Cart abandonment peaks at shipping cost reveal
TinyLlama · 1240ms · 312 tok
• Sales went down on mobile
• Desktop is better than mobile
• People leave when they see shipping
Template: Bullet Summariser v3 ✓ Saved
Observability & Logs

See exactly what every
AI agent did

Every model call across every TotalApp screen is logged automatically — tokens consumed, tools called, latency, errors. One dashboard for all your AI activity.

  • Per-event log with token breakdown, duration, tool calls, and error trace
  • Analytics view — token usage by source app, model, and user
  • Cross-application tracking — all AI screens log here automatically
  • Export logs as CSV for cost reporting and compliance audits
Open Observability
totalapp.app / orch-observability
Total Calls
2,847
Tokens Used
1.2M
Avg Latency
342ms
Error Rate
0.8%
TimeSourceModelTokensms
14:22:01 Agent Arch. claude-sonnet-4-6 1,247 284
14:21:58 Prompt Lab claude-sonnet-4-6 847 312
14:21:45 HR Reports claude-sonnet-4-6 2,103 891
14:21:30 Tool Registry tinyllama-1.1b 312 124
14:21:12 Ecommerce claude-sonnet-4-6 3,891 timeout
All Capabilities

Every AI layer, one platform

Design, run, test, monitor, and extend AI agents without leaving TotalApp.

Agent Architect

Design multi-step AI agents visually — pick LLM, wire prompt chains, attach tools, and export as reusable workflow nodes.

Model Center

Manage every local and cloud LLM from one screen. Switch the active model for any agent or workflow with a single click.

Prompt Lab

Test and refine prompts in a live sandbox. Compare outputs across models side-by-side before wiring them into production agents.

Knowledge Base

Upload PDFs and text documents, convert to vector embeddings, and make them available to any agent node as RAG context.

Agent Persona Editor

Define each agent's character, capabilities, and constraints with detailed system prompts. Give agents consistent identities.

Tool / Function Registry

Define JSON Schema-based tools — database queries, API calls, file I/O, email — and wire them into any agent with one click.

Observability & Logs

Every AI call across every TotalApp screen is logged — tokens, latency, tool calls, and errors — in one unified dashboard.

Pre-built Agent Nodes

Drop Standard, Reasoning, Autonomous, Multimodal, and Domain-specific agent nodes directly into any workflow.