DocsGetting Started

MURA SDK

Python SDK for the MURA Supply Chain Agent Network. Enable AI agents to coordinate procurement, logistics, and compliance.

For Platforms

Use MuraClient to power your procurement with MURA's agent network

For Suppliers

Use SupplierAgent to join the network and receive RFQs automatically

Built on NANDA — MURA implements concepts from MIT Media Lab's NANDA research for agent coordination.

Installation

bash

1pip install mura-sdk

Requires Python 3.10+

Quick Start

For Platforms

python

1from mura import MuraClient
2
3client = MuraClient(api_key="mura_live_xxx")
4
5result = client.procure(
6    request="500 temperature sensors",
7    budget=5000,
8    destination_region="EU"
9)
10
11print(result.recommendation.recommended_supplier)

For Suppliers

python

1from mura import SupplierAgent
2
3agent = SupplierAgent(
4    name="Acme Electronics",
5    capabilities=["electronics", "sensors"],
6    region="EU"
7)
8
9agent.add_to_catalog("temperature_sensor", 5.50, "sensors")
10agent.register()
11agent.run()

SDKMuraClient

MuraClient

The main client for platforms to use MURA as supply chain infrastructure.

python

1from mura import MuraClient
2
3client = MuraClient(
4    api_key="mura_live_xxx",      # Your API key
5    base_url="https://mura-production.up.railway.app",  # Optional
6    timeout=120.0                  # Request timeout
7)

Parameters

Name	Type	Description
api_key	`str`	Your MURA API key
base_url	`str`	API base URL (optional)
timeout	`float`	Request timeout in seconds (default: 120)

client.procure(request, budget=None, deadline_days=7, destination_region='EU') → ProcurementResult

Run the full procurement workflow: BOM → discovery → quoting → compliance → logistics → recommendation.

python

1result = client.procure(
2    request="Build me a racing drone",
3    budget=1000,
4    deadline_days=14,
5    destination_region="EU"
6)
7
8# Access results
9result.bom                    # Bill of Materials
10result.quotes                 # All quotes
11result.compliance             # Compliance result
12result.logistics              # Shipping plan
13result.recommendation         # Final recommendation

Parameters

Name	Type	Description
request*	`str`	Natural language procurement request
budget	`float`	Budget constraint in EUR
deadline_days	`int`	Delivery deadline (default: 7)
destination_region	`str`	Destination: EU, US, Asia

async client.procure_stream(request, ...) → AsyncGenerator[dict]

Stream procurement with real-time updates. Perfect for live UI progress.

python

1async for step in client.procure_stream("500 sensors"):
2    print(f"[{step['agent']}] {step['message']}")
3
4    if step.get('phase') == 'complete':
5        result = step.get('data')

Step Object

{
  "agent": "compliance",
  "message": "Checking EU regulations...",
  "phase": "compliance",
  "progress": 0.7
}

client.registry.discover(role=None, capability=None, region=None, min_trust=0.0) → List[AgentFacts]

Discover agents in the MURA network.

python

1from mura import AgentRole
2
3suppliers = client.registry.discover(
4    role=AgentRole.SUPPLIER,
5    capability="electronics",
6    region="EU",
7    min_trust=0.8
8)
9
10for s in suppliers:
11    print(f"{s.name}: {s.trust.reputation_score}")

client.quotes.get_quote(supplier_id, items, deadline_days=None) → dict

python

1quote = client.quotes.get_quote(
2    supplier_id="supplier-abc123",
3    items=[
4        {"part_name": "temperature_sensor", "quantity": 100},
5        {"part_name": "humidity_sensor", "quantity": 50}
6    ]
7)
8
9print(f"Total: €{quote['total_cost']}")

client.compliance.check(items, destination_region, transport_type='air') → dict

python

1result = client.compliance.check(
2    items=[{"part_name": "battery_li_ion", "category": "electronics"}],
3    destination_region="EU",
4    transport_type="air"
5)
6
7print(f"Status: {result['status']}")  # passed, warning, failed

client.logistics.plan(origin_region, destination_region, items, deadline_days=None) → dict

python

1plan = client.logistics.plan(
2    origin_region="Asia",
3    destination_region="EU",
4    items=[{"weight_kg": 5}],
5    deadline_days=7
6)
7
8print(f"Provider: {plan['provider']}, Cost: €{plan['shipping_cost']}")

SDKSupplierAgent

SupplierAgent

Join the MURA network as a supplier to receive RFQs and respond with quotes.

python

1from mura import SupplierAgent
2
3agent = SupplierAgent(
4    name="Acme Electronics",
5    capabilities=["electronics", "sensors", "MCUs"],
6    region="EU",
7    country="Germany",
8    certifications=[
9        {"authority": "ISO", "certification": "9001"},
10    ],
11    max_discount_pct=15.0,
12    lead_time_days=5
13)

Parameters

Name	Type	Description
name*	`str`	Your company name
capabilities*	`List[str]`	What you supply
region*	`str`	Your region: EU, US, Asia
country	`str`	Your country
certifications	`List[dict]`	Your certifications

Catalog Management

python

1# Add items to catalog
2agent.add_to_catalog(
3    part_name="temperature_sensor",
4    unit_price=5.50,
5    category="sensors",
6    stock=1000,
7    lead_time_days=3
8)
9
10# Load from JSON file
11agent.load_catalog_from_json("catalog.json")
12
13# Load from dictionary
14agent.load_catalog_from_dict({
15    "esp32": {"unit_price": 8.0, "category": "MCUs", "stock": 500}
16})

Custom RFQ Handler

Connect to your ERP/inventory system with a custom handler.

python

1from mura import Quote, QuoteItem
2
3@agent.on_rfq
4def handle_rfq(rfq):
5    items = []
6
7    for item in rfq.items:
8        price = my_erp.get_price(item.part_name)
9        if price is None:
10            continue
11
12        items.append(QuoteItem(
13            part_name=item.part_name,
14            unit_price=price,
15            quantity=item.quantity,
16            total_price=price * item.quantity,
17            lead_time_days=3,
18            in_stock=True
19        ))
20
21    return Quote(
22        supplier_id=agent.agent_id,
23        supplier_name=agent.name,
24        region=agent.region,
25        items=items,
26        total_cost=sum(i.total_price for i in items),
27        currency="EUR",
28        lead_time_days=max(i.lead_time_days for i in items)
29    )
30
31agent.register()
32agent.run()

ReferenceModels

Models

python

1from mura import (
2    # Agent Identity (NANDA)
3    AgentFacts,
4    TrustProfile,
5    Certification,
6
7    # Procurement
8    BOMItem,
9    BillOfMaterials,
10    Quote,
11    QuoteItem,
12    RFQ,
13
14    # Results
15    ProcurementResult,
16    ComplianceResult,
17    LogisticsPlan,
18    Recommendation,
19)

Enums

python

1from mura import TrustLevel, AgentRole, ComplianceStatus
2
3# Trust Levels (ZTAA)
4TrustLevel.SELF_DECLARED       # Agent claims capabilities
5TrustLevel.PEER_ATTESTED       # Other agents vouch
6TrustLevel.AUTHORITY_VERIFIED  # Official verification
7
8# Agent Roles
9AgentRole.SUPPLIER
10AgentRole.LOGISTICS
11AgentRole.COMPLIANCE
12AgentRole.BUYER
13
14# Compliance Status
15ComplianceStatus.PASSED
16ComplianceStatus.WARNING
17ComplianceStatus.FAILED

Error Handling

python

1from mura import (
2    MuraError,               # Base exception
3    MuraConnectionError,     # Network issues
4    MuraAuthenticationError, # Invalid API key
5    NoSuppliersFoundError,   # No matching suppliers
6    ComplianceError,         # Compliance blockers
7)
8
9try:
10    result = client.procure("dangerous goods")
11except ComplianceError as e:
12    print(f"Blocked: {e.blockers}")
13except NoSuppliersFoundError:
14    print("No suppliers available")