Connecting to an AI server

Connecting to an AI server involves configuring network access, using SDKs or APIs, and optionally setting up a local server for secure and high-performance AI operations.Network and Access RequirementsTo connect to an AI server, ensure your network allows outbound traffic to the server's domain and port. For example, Burp AI requires HTTPS access to ai.portswigger.net on port 443, and firewalls or content filters may need to allowlist the server domain. If your network uses an upstream or intercepting proxy, you may need to configure your client to trust the proxy's certificate or provide proxy credentials .Using an AI Server SDKMany AI servers provide SDKs for programmatic access. For instance, the ai-server-sdk Python package allows you to connect to an AI server using a ServerClient object. You provide your server URL, access key, and secret key to authenticate. Once connected, you can run queries, retrieve JSON responses, or integrate with frameworks like LangChain for advanced workflows. Example usage includes creating a ModelEngine instance and executing queries or database operations through the SDK .PythonCopyfrom ai_server import ServerClient, ModelEngineloginKeys = {"secretKey":"", "accessKey":""}server_connection = ServerClient(base=", access_key=loginKeys['accessKey'], secret_key=loginKeys['secretKey'])model = ModelEngine(engine_id="", insight_id=server_connection.cur_insight)response = model.run_query("Your query here") Setting Up a Local AI ServerFor high-performance or private AI operations, you can build a local AI server. A typical setup includes a powerful CPU (e.g., AMD Ryzen 9 7950X), large RAM (128GB DDR5), and GPUs (e.g., dual NVIDIA RTX 4090) to handle large models efficiently. Operating systems like Pop!_OS or Ubuntu are commonly used. Once installed, you can access an admin portal to configure AI providers, generate API keys, and manage requests locally .Admin Portal and API AccessAfter installation, AI servers often provide a web-based admin portal (e.g., :5006/admin) to manage API keys, configure AI models, and monitor server activity. You can make API requests directly to the server endpoints or use SDKs for programmatic access. If using GPU-based agents for image or video processing, additional Docker containers like ComfyUI may be installed and registered with the main AI server .Enterprise IntegrationFor cloud or enterprise AI services, such as NetSuite AI Connector, connecting involves logging into the AI client (e.g., Claude or ChatGPT), adding the connector, and authorizing access to your account. Subsequent connections reuse the authorization, and you can manage connectors through the client interface .Troubleshooting TipsVerify internet connectivity and server reachability by visiting the server URL in a browser.Check firewall, proxy, or VPN settings that may block traffic.Ensure SDK dependencies are installed and compatible with your Python environment.For local servers, confirm Docker and Docker Compose are installed and running.Reset configurations if necessary by clearing local data directories and re-running installation scripts . By following these steps, you can establish a secure and functional connection to an AI server, whether cloud-based or local, and begin executing queries or integrating AI capabilities into your applications.