January 15, 2026

featuresdks

Python & TypeScript SDKs

Official client libraries for TSFM.ai

#Python & TypeScript SDKs

We are standing up first-party client repositories for Python and TypeScript, with the docs now treating them as the intended source-of-truth SDK surfaces.

Current status:

  • TypeScript source repo target: https://github.com/tsfmai/tsfm-ts
  • Python source repo target: https://github.com/tsfmai/tsfm-python
  • Package publication is not live yet
  • The docs may reference private or not-yet-public source repositories while the rollout completes

#Python

git clone https://github.com/tsfmai/tsfm-python.git
cd tsfm-python
uv sync
from tsfm import TsfmClient

client = TsfmClient(api_key="your-api-key")

result = client.forecast({
    "model": "amazon/chronos-bolt-small",
    "inputs": [{
        "item_id": "series_1",
        "target": [1.2, 3.4, 5.6, 7.8, 9.0],
    }],
    "parameters": {
        "prediction_length": 24,
        "freq": "H",
    },
})
print(result["predictions"][0]["mean"])

#TypeScript

git clone https://github.com/tsfmai/tsfm-ts.git
cd tsfm-ts
npm install
import { TsfmClient } from "@tsfm/client";

const client = new TsfmClient({ apiKey: "your-api-key" });

const result = await client.forecast({
  model: "amazon/chronos-bolt-small",
  inputs: [{ item_id: "series_1", target: [1.2, 3.4, 5.6, 7.8, 9.0] }],
  parameters: { prediction_length: 24, freq: "H" },
});
console.log(result.predictions[0].mean);

#Features

Both SDKs support:

  • Schema-driven operation discovery from the TSFM OpenAPI contract
  • Thin convenience methods for forecast, model discovery, schedules, health, and credential validation
  • Shared auth and error handling without hiding the underlying API contract
  • Runnable source repos that can be used before package publication

Package registry publication will follow after the source repositories and docs rollout settle.