Skip to main content

Dispatch Engine Modes

Benchmarked against: Anthropic — Fast mode Tools: dispatch_work_order, call_model, search_web Architecture: Local UBI dispatch worker

SuperPortia's dispatch system routes Work Orders to different AI engines based on the engine parameter. Each engine operates in a specific mode — from free text-only analysis to full-capability code operations.


Engine modes overview

ModeEngineCostCapabilitiesUse case
Freegroq$0Text analysis, summariesTrivial research, cleanup
Free + Searchgroq-search$0Web search + analysisIntel gathering, research
CheapgeminiCentsGeneral tasksStandard analysis
Cheap + Searchgemini-search~$0.014Google Search Grounding + citationsAuthoritative research
CheapdeepseekCentsStrong reasoningAnalysis, logic tasks
CheapmistralCentsGeneral tasksEuropean alternative
CheapzhipuCentsChinese NLPChinese content tasks
Fullclaude~$1-2/runFile ops, code, shell, full MCPCode editing, system changes
IngestingestFreeMTAAA pipeline onlyBatch file ingestion

How dispatch works

Dispatch command

dispatch_work_order(
work_order_id="WO-2026-0305-001",
engine="groq" # or gemini, claude, etc.
)

The worker runs as a background process (fire-and-forget). Check results with check_dispatch() or get_work_order_detail().


Mode capabilities comparison

CapabilityFree (Groq)Cheap (Gemini)Full (Claude)Ingest
Read WO descriptionYesYesYesYes
Text generationYesYesYesNo
Web searchGroq-search onlyGemini-search onlyVia toolsNo
File readNoNoYesYes
File write/editNoNoYesNo
Shell commandsNoNoYesNo
UB ingestionNoNoYesYes
Code executionNoNoYesNo
Git operationsNoNoYesNo

Flat vs. nested mode

LangGraph agents can operate in two modes depending on the engine's tool-calling capability:

ModeEnginesHow it works
FlatGroq, DeepSeek, Mistral, ZhipuDirect tool invocation — simple but limited
NestedGemini, ClaudeAgent-as-tool pattern — agents call other agents

The supervisor.py automatically selects the correct mode based on the engine. This is transparent to the WO creator.

Why the split: Groq uses XML format for tool calling instead of JSON, which breaks nested agent-as-tool patterns. DeepSeek, Mistral, and Zhipu have limited tool-calling support.


Engine selection for dispatch

WO typeRecommended engineWhy
Research / intel gatheringgroq-search or gemini-searchFree/cheap, web search built in
Text analysis / summarygroq or deepseekFree/cheap, text-only task
Chinese content analysiszhipuBest Chinese NLP
Code changesclaudeOnly engine with file operations
Batch file ingestioningestMTAAA pipeline, no LLM cost
General task with quality needgeminiGood balance of cost and quality

PageRelationship
Choosing an EngineFull selection guide
Engine OverviewAll engine capabilities
Local UBI MCPDispatch worker details
WO AdminWork order system