​
Overview

In our knowledge graph service, an Environment acts as a logical container that isolates sources within a user’s workspace. This allows users to create and manage multiple knowledge graphs independently under the same account. Each environment provides powerful AI-driven search capabilities through our Intelligent Search system, enabling natural language queries, automatic analytical operations, and sophisticated graph traversal.

​
Why Use Environments?

AI agents often interact with diverse and unrelated sets of information. Environments provide a clean way to separate these contexts, ensuring that data from one use case or project does not interfere with another. This is especially useful when:
  • Domain-Specific Knowledge Graphs: Separate financial data from customer support conversations
  • Multi-Tenant Applications: Isolate data between different clients or organizations
  • Specialized AI Agents: Create focused contexts for specific agent capabilities
  • Development vs Production: Maintain separate environments for testing and live data
  • Compliance and Security: Ensure sensitive data remains isolated from general datasets

​
Key Characteristics

  • Data Isolation: Sources stored in one environment are not accessible from another, guaranteeing clean separation of information
  • Scoped Operations: All search, extraction, and graph operations are performed within the context of a specific environment
  • AI-Powered Search: Each environment includes intelligent search capabilities with natural language understanding
  • Flexible Architecture: Create unlimited environments for fine-grained control over your knowledge graphs
  • Multi-Modal Search: Support for vector search, graph traversal, type-aware queries, and analytical operations

​
Structure

Each environment has the following properties:
FieldTypeDescription
namestringUnique name of the environment (required)
descriptionstringOptional description providing context or purpose
Environment names must be unique across all of a user’s environments. Attempting to create another environment with the same name will result in an error.

​
Relationship with Sources

Every source must be associated with exactly one environment. This ensures that all knowledge derived from a source is scoped correctly and remains isolated within its intended graph.

​
Supported Source Types

  • Conversations: Chat messages, dialogues, and communication threads
  • Business Data: JSON datasets, customer records, financial data
  • Files: PDF documents, structured data files
  • NetworkX Graphs: Direct graph imports for complex relationship data

​
Search and Analytics Capabilities

Each environment provides advanced search functionality: 
from praxos_python import SyncClient

client = SyncClient(api_key="prx_your_api_key_here")
env = client.get_environment(name="financial_analysis")

# Natural language queries with AI analysis
results = env.intelligent_search("What are the highest risk transactions this quarter?")

# Automatic analytical operations  
results = env.intelligent_search("Calculate average deal size by customer tier")

# Graph-aware relationship queries
results = env.intelligent_search("Find all entities connected to suspicious accounts")

​
Example Use Cases

​
Financial Services Environment

# Create specialized environment for financial data
financial_env = client.create_environment(
    name="financial_compliance",
    description="Financial transactions and compliance monitoring"
)

# Query with domain-specific intelligence
suspicious_transactions = financial_env.intelligent_search(
    "Find transactions above $10k with high risk indicators"
)

​
Customer Support Environment

# Separate environment for customer interactions
support_env = client.create_environment(
    name="customer_support", 
    description="Customer conversations and issue tracking"
)

# Analyze support patterns
issue_analysis = support_env.intelligent_search(
    "What are the most common customer complaints this month?"
)