Timesfm Forecasting

A zero-shot time series forecaster built on Google's TimesFM foundation model. Feed it any univariate series (sales, sensor data, stock prices) and get point forecasts plus calibrated prediction intervals without training anything. The 200M parameter model needs about 4GB RAM and includes a mandatory preflight checker to verify your machine can handle it before loading. Supports basic forecasting and covariate-based forecasting (XReg) with exogenous variables. The honest take: this is genuinely useful if you need quick forecasts across many different series without tuning ARIMA parameters for each one, but you're still loading an 800MB model into memory. Works with context lengths up to 16,384 points and handles batch forecasting efficiently.

skills.sh installs: 187
GitHub stars: 35.4K
last updated: Jun 3, 2026

View on GitHub

Install

$ npx -y skills add google-research/timesfm --skill timesfm-forecasting --agent claude-code

Installs into .claude/skills of the current project.

Files

TimesFM Forecasting

Overview

TimesFM (Time Series Foundation Model) is a pretrained decoder-only foundation model developed by Google Research for time-series forecasting. It works zero-shot — feed it any univariate time series and it returns point forecasts with calibrated quantile prediction intervals, no training required.

This skill includes a mandatory preflight system checker that verifies RAM, GPU memory, and disk space before the model is ever loaded so the agent never crashes the user's machine.

Key numbers: TimesFM 2.5 uses 200M parameters (~800 MB on disk, ~1.5 GB in RAM on CPU, ~1 GB VRAM on GPU). The archived v1/v2 500M-parameter model needs ~32 GB RAM. Always run the system checker first.