Skip to main content

​
Persistence Quickstart

Enable automatic conversation persistence with memory configuration.

​
Installation

pip install "praisonaiagents[tools]"

​
Docker Setup

# PostgreSQL
docker run -d --name praison-postgres -p 5432:5432 \
    -e POSTGRES_PASSWORD=praison123 \
    -e POSTGRES_DB=praisonai \
    postgres:16

# Qdrant (optional - for knowledge)
docker run -d --name praison-qdrant -p 6333:6333 qdrant/qdrant

# Redis (optional - for state)
docker run -d --name praison-redis -p 6379:6379 redis:7

​
Basic Usage

from praisonaiagents import Agent

# Create agent with persistence
agent = Agent(
    name="Assistant",
    instructions="You are a helpful assistant.",
    memory={
        "db": "postgresql://postgres:praison123@localhost:5432/praisonai",
        "session_id": "my-session-001"
    }
)

# All chats are automatically persisted
response = agent.start("My name is Alice")
print(response)

​
SQLite (Zero Config)

from praisonaiagents import Agent

agent = Agent(
    name="Assistant",
    memory={
        "db": "conversations.db",
        "session_id": "local-session"
    }
)
response = agent.start("Hello!")
print(response)

​
Backend Options

from praisonaiagents import Agent

# PostgreSQL
agent = Agent(
    name="Bot",
    memory={"db": "postgresql://user:pass@localhost/db"}
)

# SQLite
agent = Agent(
    name="Bot",
    memory={"db": "mydata.db"}
)

# MySQL
agent = Agent(
    name="Bot",
    memory={"db": "mysql://user:pass@localhost/db"}
)

​
Environment Variables

export PRAISON_CONVERSATION_URL="postgresql://localhost/mydb"
export OPENAI_API_KEY="your-key"

import os
from praisonaiagents import Agent

agent = Agent(
    name="Assistant",
    memory={"db": os.getenv("PRAISON_CONVERSATION_URL")}
)

​
CLI Usage

# Validate connectivity
praisonai persistence doctor \
    --conversation-url "postgresql://postgres:praison123@localhost/praisonai"

# Run agent with persistence
praisonai persistence run \
    --session-id "cli-session" \
    --conversation-url "postgresql://postgres:praison123@localhost/praisonai" \
    "Hello, my name is Bob"

​
Continue a Session from the CLI

praisonai session resume restores chat history, model, and agent name — then optionally continues with a new prompt: 
praisonai session resume cli-session "What was my name again?"

​
Serverless Database Tuning

For serverless databases that may experience cold starts, tune the retry parameters: 
from praisonaiagents import Agent

# PlanetScale MySQL 
agent = Agent(
    name="Assistant",
    memory={
        "db": "mysql://user:pass@gateway.psdb.cloud:3306/mydb",
        "max_retries": 5,      # Higher retries for cold starts
        "retry_delay": 1.0     # Longer initial delay
    }
)

# Neon PostgreSQL
agent = Agent(
    name="Assistant", 
    memory={
        "db": "postgresql://user:pass@host.neon.tech/db",
        "max_retries": 3,
        "retry_delay": 0.5
    }
)
Supported serverless platforms with auto-retry:
  • PlanetScale: .psdb.cloud hosts (exponential backoff)
  • Neon: .neon.tech hosts
  • CockroachDB: .cockroachlabs.cloud, .cockroachlabs.com
  • Supabase: .supabase.com, .supabase.co
  • Xata: .xata.sh

​
Next Steps