timesfm
by google-research
TimesFM is a pretrained model from Google Research that predicts future values in time-series data, such as sales, demand or traffic figures. It can be run from Python, and Google also offers it through BigQuery ML and Google Sheets.
- Stars
- 34.8k
- Licence
- Apache-2.0
- Runs on
- Your computer
- Setup
- Technical
- Language
- Python
- Age
- 913 days
What you can do with it
- Forecast future sales, demand or traffic from past figures without training a model first
- Forecast several related data series together, using extra inputs such as promotions or holidays
- Fine-tune the model on your own data using the included LoRA example
Make money with it
A freelance analyst could offer forecasting reports to small businesses, using the Apache-2.0 version 2.5 weights or Google's authorised cloud services for commercial work.
An idea, not financial advice. Check the licence before charging for anything built on it.
Ways a business could use it
- Forecast stock levels or demand to cut over-ordering and stock-outs
- Predict future revenue or website traffic to plan budgets and staffing
Can you use it commercially?
Free to use, change and sell, including in paid products. Keep the copyright notice.
Not legal advice. Read the project's licence file before building a product on it.
Getting it running
Install the Python package with pip (pip install timesfm[torch], or timesfm[mlx] on Apple silicon) and download the model weights from Hugging Face.
Follow the official instructions in the project README.
Things to check
- TimesFM 3.0 weights are under a non-commercial licence, so commercial or production use is not allowed on self-hosted copies
- Commercial use of 3.0 is only permitted through Google Cloud services such as BigQuery ML, which are paid
- The open version is not an officially supported Google product
- A GPU is likely to be needed for fast inference on large workloads
Who's talking about it
Checked 29 Oct 2026. Summary drafted by AI from the project's README.
