Skip to main content

​
Persistence Module

The Persistence module provides database adapters and session management for storing agent conversations, runs, and traces.

​
Installation

pip install praisonai

# For specific database support
pip install "praisonai[postgres]"
pip install "praisonai[redis]"

​
Quick Start

The recommended way to enable persistence is via the memory={} parameter: 
from praisonaiagents import Agent

agent = Agent(
    name="Assistant",
    instructions="You are helpful.",
    memory={
        "db": "postgresql://localhost/mydb",
        "session_id": "my-session"
    }
)

response = agent.start("Hello!")

​
Supported Databases

​
Conversation Stores

DatabaseInstallURL Format
PostgreSQLpip install "praisonai[postgres]"postgresql://user:pass@host/db
MySQLpip install "praisonai[mysql]"mysql://user:pass@host/db
SQLiteBuilt-insqlite:///path/to/db.sqlite
JSONBuilt-injson:///path/to/data.json

​
Knowledge Stores (Vector)

DatabaseInstall
Qdrantpip install "praisonai[qdrant]"
Chromapip install "praisonai[chroma]"
Pineconepip install "praisonai[pinecone]"
PGVectorpip install "praisonai[pgvector]"

​
State Stores

DatabaseInstall
Redispip install "praisonai[redis]"
MongoDBpip install "praisonai[mongodb]"

​
Database Examples

​
PostgreSQL

from praisonaiagents import Agent

agent = Agent(
    name="Assistant",
    memory={
        "db": "postgresql://user:pass@localhost/mydb",
        "session_id": "user-123"
    }
)

​
SQLite

from praisonaiagents import Agent

agent = Agent(
    name="Assistant",
    memory={
        "db": "sqlite:///./agent.db",
        "session_id": "local-session"
    }
)

​
Redis

from praisonaiagents import Agent

agent = Agent(
    name="Assistant",
    memory={
        "backend": "redis",
        "db": "redis://localhost:6379",
        "session_id": "redis-session"
    }
)

​
Session Management

Sessions allow you to maintain conversational context: 
from praisonaiagents import Agent

# Create agent with session
agent = Agent(
    name="Assistant",
    memory={"db": "sqlite:///data.db", "session_id": "user-123"}
)

# Conversations persist across runs
response1 = agent.start("Remember my name is Alice")
# Later...
response2 = agent.start("What's my name?")  # Remembers "Alice"

​
Data Models

​
DbMessage

Stored message format: 
{
    "id": "msg-uuid",
    "session_id": "session-123",
    "agent_id": "agent-name",
    "role": "user",  # or "assistant"
    "content": "Message content",
    "timestamp": "2024-01-01T00:00:00Z"
}

​
DbRun

Execution run record: 
{
    "id": "run-uuid",
    "session_id": "session-123",
    "agent_id": "agent-name",
    "status": "completed",
    "started_at": "2024-01-01T00:00:00Z",
    "completed_at": "2024-01-01T00:00:01Z"
}

​
DbToolCall

Tool execution record: 
{
    "id": "tool-uuid",
    "run_id": "run-uuid",
    "tool_name": "search",
    "arguments": {"query": "AI news"},
    "result": "Search results...",
    "timestamp": "2024-01-01T00:00:00Z"
}

​
See Also