tg-start-flow

Starts a processing flow using a defined flow blueprint.

Synopsis

tg-start-flow -n CLASS_NAME -i FLOW_ID -d DESCRIPTION [options]

Description

The tg-start-flow command creates and starts a new processing flow instance based on a predefined flow blueprint. Flow blueprints define the processing pipeline configuration, while flow instances are running implementations with specific identifiers.

New in v1.4: Flows can be customized with configurable parameters that control LLM models, chunking behavior, and other processing settings.

Options

Required Arguments

Option	Description
`-n, --blueprint-name CLASS`	Name of the flow blueprint to instantiate
`-i, --flow-id FLOW_ID`	Unique identifier for the new flow instance
`-d, --description DESC`	Human-readable description of the flow

Optional Arguments

Option	Default	Description
`-u, --api-url URL`	`$TRUSTGRAPH_URL` or `http://localhost:8088/`	TrustGraph API URL
`-t, --token TOKEN`	`$TRUSTGRAPH_TOKEN`	Authentication token

Flow Parameters (New in v1.4)

Option	Description
`-p, --parameters JSON`	Flow parameters as JSON string
`--parameters-file FILE`	Path to JSON file containing flow parameters
`--param KEY=VALUE`	Individual parameter (can be used multiple times)

Note: All parameter values are stored as strings internally.

Flow blueprints are free to define any parameters they need. The standard flow blueprints provided by TrustGraph (everything, structured, ontology) support the following parameters:

Parameter	Description	Notes
`llm-model`	LLM model to use
`llm-rag-model`	LLM model for RAG operations	Optional, falls back to `llm-model` if not specified
`llm-temperature`	LLM temperature
`llm-rag-temperature`	LLM temperature for RAG operations	Optional, falls back to `llm-temperature` if not specified
`embeddings-model`	Embeddings model
`chunk-size`	Document chunk size
`chunk-overlap`	Document chunk overlap

All parameters have usable defaults if not specified.

Examples

Start Basic Flow

tg-start-flow \
  -n "everything" \
  -i "research-flow" \
  -d "Research document processing pipeline"

Start with Parameters

tg-start-flow \
  -n "everything" \
  -i "custom-flow" \
  -d "Custom flow with parameters" \
  --param llm-model=claude-sonnet-4-20250514 \
  --param llm-temperature=0.7 \
  --param chunk-size=512

Using Parameters File

tg-start-flow \
  -n "ontology" \
  -i "ontology-flow" \
  -d "Ontology RAG processing" \
  --parameters-file flow-config.json

Using JSON Parameters

tg-start-flow \
  -n "structured" \
  -i "prod-flow" \
  -d "Structured data extraction" \
  -p '{"llm-model": "claude-sonnet-4-20250514", "llm-temperature": "0.7"}'

Environment Variables

TRUSTGRAPH_URL: Default API URL
TRUSTGRAPH_TOKEN: Default authentication token

tg-stop-flow - Stop a running flow
tg-show-flows - List all running flows
tg-show-flow-blueprints - List available blueprints

API Integration

This command uses the Flow Management API to create and start flow instances from blueprint definitions.