quartermaster-graph 0.6.3

Framework-agnostic agent graph schema for AI agent workflows as directed acyclic graphs

quartermaster-graph

Framework-agnostic graph schema for defining AI agent workflows as directed graphs.

Features

• Pydantic-based models: Agent, GraphSpec, GraphNode, GraphEdge with full validation
• GraphBuilder fluent API for programmatic graph construction (Graph convenience alias)
• 40+ node types covering LLM, control flow, data, user interaction, memory, and utility
• YAML/JSON serialization with round-trip fidelity
• Graph validation: start/end nodes, cycle detection, orphan detection, edge label checks
• Graph traversal: topological sort, path finding, successor/predecessor queries
• Pre-built templates: simple chat, decision tree, multi-agent supervisor, and more
• Typed metadata schemas for each node type

New in v0.5.0

• program_runner(program=<callable>) -- pass a @tool()-decorated function directly instead of its name string; the graph builder auto-registers it. Parity with .agent(tools=[...]).

Installation

pip install quartermaster-graph

Dependencies: pydantic>=2.0, pyyaml>=6.0

Quick Start

Build a Graph with the Fluent API

from quartermaster_graph import Graph

graph = (
    Graph("Customer Support Agent")
    .user("How can I help you?")
    .instruction("Classify intent", model="gpt-4o", system_instruction="Classify the user's intent.")
    .decision("Route by intent", options=["billing", "technical", "general"])
    .on("billing").instruction("Handle billing", model="gpt-4o").end()
    .on("technical").instruction("Handle technical", model="gpt-4o").end()
    .on("general").instruction("Handle general", model="gpt-4o").end()
    .end()
)

Graph is a convenience alias for GraphBuilder. Both .build() and .to_graph() return a GraphSpec:

from quartermaster_graph import GraphBuilder

builder = GraphBuilder("My Agent")
builder.instruction("Process", model="gpt-4o").end()

spec = builder.build()       # returns GraphSpec
spec = builder.to_graph()    # same thing

Load a Graph from YAML

# agent.yaml
agent_id: "550e8400-e29b-41d4-a716-446655440000"
start_node_id: "00000000-0000-0000-0000-000000000001"
nodes:
  - id: "00000000-0000-0000-0000-000000000001"
    type: "Start1"
    name: "Start"
  - id: "00000000-0000-0000-0000-000000000002"
    type: "User1"
    name: "Input"
  - id: "00000000-0000-0000-0000-000000000003"
    type: "Instruction1"
    name: "Process"
    metadata:
      llm_system_instruction: "Analyze user input"
      llm_model: "gpt-4o"
  - id: "00000000-0000-0000-0000-000000000004"
    type: "End1"
    name: "End"
edges:
  - source_id: "00000000-0000-0000-0000-000000000001"
    target_id: "00000000-0000-0000-0000-000000000002"
  - source_id: "00000000-0000-0000-0000-000000000002"
    target_id: "00000000-0000-0000-0000-000000000003"
  - source_id: "00000000-0000-0000-0000-000000000003"
    target_id: "00000000-0000-0000-0000-000000000004"

from quartermaster_graph import from_yaml

with open("agent.yaml") as f:
    spec = from_yaml(f.read())

API Reference

Core Models

Model	Description
Agent	Top-level agent definition (name, description, tags)
GraphSpec	Graph definition: nodes, edges, start node, features
GraphNode	A node with type, metadata, traversal config, error handling
GraphEdge	Directed edge with optional label and routing points
NodePosition	Visual position for editor rendering

AgentGraph and AgentVersion both exist as deprecated backward-compatibility aliases for GraphSpec.

GraphSpec Methods

Method	Return Type	Description
get_node(node_id)	GraphNode | None	Find a node by ID
get_start_node()	GraphNode | None	Get the start node
get_successors(node_id)	list[GraphNode]	All successor nodes
get_predecessors(node_id)	list[GraphNode]	All predecessor nodes
get_edges_from(node_id)	list[GraphEdge]	All edges from a node
get_edges_to(node_id)	list[GraphEdge]	All edges to a node

GraphBuilder Methods

Start / End

Method	Returns	Description
.start()	GraphBuilder	Add a Start node (auto-inserted since v0.2.0 — rarely needed explicitly)
.end()	GraphBuilder	Add an End node (or close a branch)

LLM Nodes

Method	Returns	Description
.instruction(name, model, provider, temperature, system_instruction, **kwargs)	GraphBuilder	LLM text generation, no tools, streams response
.decision(name, model, provider, temperature, prefix_message, suffix_message, options, **kwargs)	GraphBuilder	LLM picks one path via pick_path tool (non-streaming)

| .summarize(name, model, provider, temperature, system_instruction, **kwargs) | GraphBuilder | LLM condenses conversation history | | .agent(name, model, provider, system_instruction, tools, max_iterations, **kwargs) | GraphBuilder | Agentic loop with tools, up to max_iterations | | .vision(name, model, provider, system_instruction, **kwargs) | GraphBuilder | Image vision/analysis node | | .merge(name, model, provider, temperature, system_instruction, prefix_message, suffix_message, **kwargs) | GraphBuilder | LLM merges parallel branch outputs |

Control Flow

Method	Returns	Description
.on(label)	BranchBuilder	Start a named branch for a decision/if/switch
.if_node(name, expression)	GraphBuilder	Conditional branch via expression, no LLM
.static_decision(name, expression)	GraphBuilder	Expression-based branching, no LLM
.user_decision(name)	GraphBuilder	User picks which path to follow
.switch(name, cases, default_edge_id)	GraphBuilder	Multi-way switch, first matching case wins
.parallel(name)	GraphBuilder	Start a parallel fan-out
.branch()	BranchBuilder	Start a parallel branch
.static_merge(name, text)	GraphBuilder	Merge parallel branches without LLM
.break_node(name, targets)	GraphBuilder	Stop backward message collection

User Interaction

Method	Returns	Description
.user(name, prompts)	GraphBuilder	Pause flow for user input
.user_form(name, parameters)	GraphBuilder	Show a structured form to the user

Data Nodes

Method	Returns	Description
.static(name, text)	GraphBuilder	Fixed text content, no LLM
.code(name, code, filename)	GraphBuilder	Code execution node
.text(name, template)	GraphBuilder	Jinja2 template rendering
.var(name, variable, expression)	GraphBuilder	Evaluate expression, store as variable
.text_to_variable(name, variable, source)	GraphBuilder	Convert text output to a variable
.program_runner(name, program, **kwargs)	GraphBuilder	Run a program/tool inline

Memory

Method	Returns	Description
.read_memory(name, memory_name, memory_type, variable_names)	GraphBuilder	Read from persistent memory
.write_memory(name, memory_name, memory_type, variables)	GraphBuilder	Write to persistent memory
.update_memory(name, memory_name, memory_type, variables)	GraphBuilder	Update existing memory variables
.flow_memory(name, memory_name, initial_data)	GraphBuilder	Define flow-scoped memory
.user_memory(name, memory_name, initial_data)	GraphBuilder	Define user-scoped persistent memory

Composition / Utility

Method	Returns	Description
.sub_agent(name, graph_id)	GraphBuilder	Call another agent graph synchronously
.use(sub_graph)	GraphBuilder	Inline a sub-graph (accepts GraphSpec or GraphBuilder)
.comment(name, text)	GraphBuilder	Documentation-only node, no runtime logic
.allowed_agents(*agent_ids)	GraphBuilder	Restrict which sub-agents can be spawned
.node(node_type, name, metadata, **kwargs)	GraphBuilder	Add any node type generically
.edge(source_id, target_id, label, is_main)	GraphBuilder	Manually add an edge
.connect(from_name, to_name, label)	GraphBuilder	Create an edge between two nodes by name

Build / Export

Method	Returns	Description
.build(validate=True)	GraphSpec	Build the graph, optionally validate
.to_graph(validate=True, agent_id=None)	GraphSpec	Build with optional explicit agent ID
.to_agent(validate=True)	Agent	Export as a full Agent model

Node Configuration Kwargs

All builder methods accept optional keyword arguments:

Kwarg	Type	Description
traverse_in	TraverseIn	When to execute: AWAIT_FIRST or AWAIT_ALL
traverse_out	TraverseOut	Which successors to trigger
thought_type	ThoughtType	How to build conversation context
message_type	MessageType	What message role to use
show_output	bool	Whether to display this node's output (default True)
error_handling	ErrorStrategy	What to do on failure

Loops and Cycles

Use connect() to create back-edges for iterative flows:

agent = (
    Graph("Refiner")
    .user("Input")
    .var("Init", variable="round", expression="1")
    .text("Header", template="Round {{round}}", traverse_in=TraverseIn.AWAIT_FIRST)
    .instruction("Process", system_instruction="Improve the text")
    .var("Increment", variable="round", expression="round + 1")
    .if_node("Done?", expression="round > 3")
    .on("true").text("Done", template="Complete").end()
    .on("false").text("Continue", template="", show_output=False).end()
    .end()
)
agent.connect("Continue", "Header", label="loop")
graph = agent.build(validate=False)  # skip validation for intentional cycles

Validation

from quartermaster_graph import validate_graph

errors = validate_graph(agent_graph)
for err in errors:
    print(f"[{err.severity}] {err.code}: {err.message}")

Validation checks:

• Exactly one Start node, at least one End node
• All edge source/target IDs reference existing nodes
• Orphan detection (unreachable from Start)
• Cycle detection (DAG property)
• Decision/If/Switch nodes have proper edge labels

Serialization

from quartermaster_graph import to_json, from_json, to_yaml, from_yaml, json_schema

# JSON round-trip
data = to_json(agent_graph)
restored = from_json(data)

# YAML round-trip
yaml_str = to_yaml(agent_graph)
restored = from_yaml(yaml_str)

# JSON Schema for cross-language validation
schema = json_schema()

Traversal Utilities

from quartermaster_graph import (
    get_start_node, get_successors, get_predecessors,
    get_path, topological_sort, find_merge_points, find_decision_points,
)

start = get_start_node(agent_graph)
ordered = topological_sort(agent_graph)
path = get_path(agent_graph, start_id, end_id)

Enums

Enum	Values
NodeType	40+ types: START, END, INSTRUCTION, DECISION, IF, SWITCH, AGENT, USER, USER_FORM, USER_DECISION, CODE, MERGE, STATIC, STATIC_MERGE, STATIC_DECISION, VAR, TEXT, SUMMARIZE, SUB_ASSISTANT, BREAK, COMMENT, etc.
TraverseIn	AWAIT_ALL, AWAIT_FIRST
TraverseOut	SPAWN_ALL, SPAWN_NONE, SPAWN_START, SPAWN_PICKED
ThoughtType	SKIP, NEW, NEW_HIDDEN, NEW_COLLAPSED, INHERIT, CONTINUE, EDIT_OR_NEW, EDIT_SAME, APPEND, USE_PREVIOUS, etc.
MessageType	AUTOMATIC, USER, ASSISTANT, SYSTEM, TOOL, VARIABLE
ErrorStrategy	STOP, RETRY, SKIP, CONTINUE, CUSTOM
ExceptionResolution	RETRY, BREAK, CONTINUE

Integration with Sibling Packages

# Build a graph (quartermaster-graph)
from quartermaster_graph import Graph

graph = Graph("Agent").user("Input").instruction("Process").end()

# Execute it (quartermaster-engine)
from quartermaster_engine import FlowRunner

runner = FlowRunner(graph=graph.build(), node_registry=registry)
result = runner.run("Hello!")

Contributing

See CONTRIBUTING.md for guidelines.

License

Apache License 2.0 -- see LICENSE for details.