Orchestration Engine

Intelligent Task Orchestration

Agent doesn't follow fixed scripts. It decomposes tasks on its own, runs multiple steps concurrently, and auto-adjusts plans when problems arise — like a thinking executor, not a button-pressing robot.

Intelligent Decision

Auto-Routing

A fast LLM classifier analyzes each incoming query and automatically routes it to the optimal execution mode — ReAct (Standard) for simple tasks, or DAG (Planning) for complex multi-step workflows. This is the default mode.

Frontend 3-way toggle: Auto / Standard / Planner
Auto mode uses a lightweight LLM to classify query complexity
Simple queries route to ReAct for fast, low-overhead responses
Complex queries route to DAG for parallel, multi-step execution
Dynamic Router
LATENCY: 45ms
User Query
Standard Mode
Low complexity, sequential
Planning Mode
High complexity, parallel

Standard Mode (ReAct)

Agent thinks and acts step by step. After each step, it decides what to do next based on results. Ideal for tasks requiring judgment and trial-and-error, like alert investigation, document analysis, and information retrieval.

Supports native Function Calling (OpenAI-style) and JSON-mode reasoning
Auto-recovers from errors
Send supplementary messages during execution, injected at the next step
Reasoning Trace
Thought
Checking regional inventory levels for 'Skynet X1' in SAP...
Action
erp.check_stock(item='Skynet X1', region='SH')
Observation
Stock: 42 units available in Warehouse SH-01.
workflow · ticket-triagev1.2yesTriggeron_eventLLM_AnalysisclassifySQL_Queryfetch rowsOutputnotifyifscore>.85 nodes · 5 links · saved

Planning Mode (DAG)

Agent first creates a complete execution plan, decomposing tasks into steps and identifying dependencies — independent steps run concurrently, dependent steps run sequentially. After completion, it checks results and auto-retries with adjusted plans if goals aren't met.

Auto re-planning: adjusts approach when goals aren't met, up to 3 rounds
Each step runs a full think-act loop internally
Interactive flowchart showing real-time node status, dependencies, and execution time
Multi-turn conversation support with automatic history context management

Detailed Capability Matrix

Comparison between Standard ReAct and Advanced Planning modes.

Standard ModePlanning ModeAuto-Routing
How it worksStep-by-step thinking and actingPlan first, then parallel executionLLM classifies and routes automatically
Best forTasks requiring judgment and trial-and-errorMulti-step, parallelizable tasksMixed workloads (default mode)
Self-correctionAuto-recovers from errorsAuto re-plans when goals aren't metInherits from routed mode
VisualizationStep-by-step reasoning displayInteractive flowchart with real-time statusMode indicator + routed mode visualization

Visual Pipeline (Blueprint Mode)

A full visual workflow editor with 25 node types. Design agent pipelines with drag-and-drop, supporting three progressive levels:

1

Fully Static: Like Dify workflows — each node performs a fixed operation with deterministic I/O connections

2

Semi-Dynamic: Key nodes upgrade to Agent mode, reasoning autonomously within predefined tools and knowledge bases

3

Fully Dynamic: No static pipeline, Agent plans and executes completely autonomously

"Built-in scenario templates (contract review, financial reconciliation, approval assistance, etc.), forkable and customizable."
Multi-Tier Model Stack
Reasoning Tiero1 / o3-mini

Complex DAG planning & verification

General TierGPT-4o / Claude 3.5

Standard ReAct execution & Tool-use

Fast TierGPT-4o-mini / DeepSeek

Routing, Summarization & Simple tasks

Developers

Explore our Source Available code on GitHub, contribute to the connector ecosystem, or integrate FIM One into your own applications.

git clone https://github.com/fim-ai/fim-one.git && ./start.sh

Enterprise

Need private deployment, custom connectors, or professional support? Our team is ready to help you scale your AI transformation.

Private Deploy & Isolation
SSO & Audit Logs
1-on-1 Dedicated Support
SLA Availability Guarantee