Python & TypeScript SDKs

Python & TypeScript SDKs

We are shipping official client libraries for Python and TypeScript, providing type-safe wrappers around all TSFM.ai API endpoints.

Python

pip install tsfm-ai

from tsfm_ai import TSFMClient

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

# Single forecast
result = client.forecast(
    model="chronos-bolt-small",
    values=[1.2, 3.4, 5.6, 7.8, 9.0],
    freq="h",
    prediction_length=24,
)
print(result.mean)

# List available models
models = client.list_models()
for m in models:
    print(f"{m.id}: {m.description}")

TypeScript

npm install @tsfm/client

import { TSFMClient } from "@tsfm/client";

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

// Single forecast
const result = await client.forecast({
  model: "chronos-bolt-small",
  data: { values: [1.2, 3.4, 5.6, 7.8, 9.0], freq: "h" },
  predictionLength: 24,
});
console.log(result.mean);

// Batch forecast
const batch = await client.forecastBatch({
  requests: [
    { model: "chronos-bolt-small", data: { values: [...], freq: "h" }, predictionLength: 24 },
    { model: "timesfm-2.0", data: { values: [...], freq: "D" }, predictionLength: 7 },
  ],
});

Features

Both SDKs support:

• All API endpoints — forecast, batch forecast, schedules, and model listing
• Full type safety — Pydantic models in Python, TypeScript interfaces in TS
• Automatic retries — Configurable retry logic with exponential backoff
• Streaming — For large batch responses
• Async support — AsyncTSFMClient in Python, native Promises in TypeScript

The SDKs are open source and available on PyPI and npm.