Show HN: Python SDK – forecasting with foundation time-series and tabular models

Summary

FAIM SDK Production-ready Python SDK for FAIM (Foundation AI Models) - a unified platform for time-series forecasting and tabular inference powered by foundation models. Features 🚀 Multiple Foundation Models : Time-Series : FlowState, Amazon Chronos 2.0, TiRex Tabular : LimiX (classification & regression) : 🔒 Type-Safe API : Full type hints with Pydantic validation : Full type hints with Pydantic validation ⚡ High Performance : Optimized Apache Arrow serialization with zero-copy operations : Optimized Apache Arrow serialization with zero-copy operations 🎯 Probabilistic & Deterministic : Point forecasts, quantiles, samples, and probabilistic predictions : Point forecasts, quantiles, samples, and probabilistic predictions 🔄 Async Support : Built-in async/await support for concurrent requests : Built-in async/await support for concurrent requests 📊 Rich Error Handling : Machine-readable error codes with detailed diagnostics : Machine-readable error codes with detailed diagnostics 🧪 Battle-Tested: Production-ready with comprehensive error handling Installation pip install faim-sdk Authentication Get your API key at https://faim.it.com/ from faim_sdk import ForecastClient # Initialize client with your API key client = ForecastClient ( api_key = "your-api-key" ) Quick Start import numpy as np from faim_sdk import ForecastClient , Chronos2ForecastRequest # Initialize client client = ForecastClient ( api_key = "your-api-key" ) # Prepare your time-series data # Shape: (batch_size, sequence_length, features) data = np . random . randn ( 32 , 100 , 1 ). astype ( np . float32 ) # Create probabilistic forecast request request = Chronos2ForecastRequest ( x = data , horizon = 24 , # Forecast 24 steps ahead output_type = "quantiles" , quantiles = [ 0.1 , 0.5 , 0.9 ] # 10th, 50th (median), 90th percentiles ) # Generate forecast - model inferred automatically from request type response = client . forecast ( request ) # Access predictions print ( response . quantiles . shape ) # (32, ...