# PsyNova - SDK
{% embed url="" %}
***
## Technical Requirements
* Python 3.10 or higher
* Dependencies:
* requests 2.31.0
***
## Quick Start
### Installation
```
pip install psynova==1.2
```
***
### Basic Usage
```
from psynova import PsynovaClient
# Initialize the client
client = PsynovaClient(
private_key="your_private_key",
base_url="https://api.psynova.ai" # Optional: defaults to dev environment
)
# Create a new agent
agent_id = client.create_agent(
name="My AI Agent",
version_id=1,
using_https=True,
using_socket=True
)
# Chat with the agent
response = client.chat(
agent_id=agent_id,
message="Hello, AI agent!"
)
print(response)
```
***
### Error Handling
The framework includes robust error handling through the `PsynovaAPIError` class, which provides detailed error messages and optional additional error context when API requests fail.
```
try:
response = client.chat(agent_id=1, message="Hello")
except PsynovaAPIError as e:
print(f"Error: {str(e)}")
if e.errors:
print(f"Additional details: {e.errors}")
```
***
### API Reference
#### PsynovaClient
**Constructor**
```
client = PsynovaClient(private_key, base_url="https://api.psynova.ai", timeout=30)
```
* `private_key` (str): Private key for authentication
* `base_url` (str, optional): Base URL for API. Defaults to development environment
* `timeout` (int, optional): Request timeout in seconds. Defaults to 30
**Methods**
> **get\_agent(agent\_id)**
Retrieves details of a specific agent.
```
agent = client.get_agent(agent_id=1)
```
Returns: Dictionary containing agent details (id, name, version\_id, status)
> **list\_agents()**
Lists all agents associated with the account.
```
agents = client.list_agents()
```
Returns: List of dictionaries containing agent details
> **chat(agent\_id, message)**
Sends a chat message to an agent.
```
response = client.chat(agent_id=1, message="Hello!")
```
Returns: String containing the agent's response
> **create\_agent(name, version\_id, using\_https=True, using\_socket=True)**
Creates a new agent with specified configuration.
```
agent_id = client.create_agent(
name="My Agent",
version_id=1,
using_https=True,
using_socket=True
)
```
Returns: Integer ID of the newly created agent
