Initial org card

README.md

@@ -1,10 +1,90 @@
 ---
-title: README
+title: TSFM.ai
 emoji: 📈
-colorFrom: yellow
-colorTo: green
+colorFrom: indigo
+colorTo: purple
 sdk: static
 pinned: false
+short_description: Time-series foundation models as a service.
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+<div align="center">
+
+# TSFM.ai
+
+**Time-series foundation models as a service.**
+
+One API. 49+ pretrained models. Zero fine-tuning required.
+
+[tsfm.ai](https://tsfm.ai) · [docs](https://tsfm.ai/docs) · [API reference](https://tsfm.ai/docs/api) · [pricing](https://tsfm.ai/pricing)
+
+</div>
+
+---
+
+## What we do
+
+TSFM.ai hosts the largest catalog of **pretrained time-series foundation models** behind a single, consistent inference API. Bring a history — numbers, timestamps, optional covariates — and get a calibrated forecast with quantile bands, without training or fine-tuning.
+
+We serve every major open-weights forecaster:
+
+- **Chronos / Chronos-Bolt / Chronos-2** (Amazon)
+- **TimesFM 2.0 / 2.5** (Google)
+- **Moirai 1.x / 2.0 / Moirai-MoE** (Salesforce)
+- **Granite TTM / PatchTST / FlowState** (IBM Research)
+- **TiRex** (NX-AI)
+- **Toto** (Datadog)
+- **TimeMoE** (Maple / Tsinghua)
+- **MOMENT** (AutonLab, CMU)
+- **Sundial / Timer / TimeMixer++** (THUML)
+- **Lag-Llama**, **TEMPO**, **Kairos**, **YingLong**, **Kronos**, **TSPulse** and more.
+
+If the model is on the Hub and has weights, we're probably already serving it. If not, open an issue on the repo and we'll prioritize it.
+
+## Why time-series deserves its own provider
+
+General-purpose LLM inference stacks are a bad fit for forecasting. Time-series models have narrow context windows, variable history lengths, quantile outputs, exogenous covariates, and probabilistic sampling — none of which map cleanly onto OpenAI-style APIs. We built TSFM.ai specifically for this surface: our request/response schema carries `past_values`, `past_timestamps`, `past_covariates`, `future_covariates`, `static_covariates`, `quantiles`, and `num_samples` as first-class fields. Batch forecasting, ensemble routing, and MCP tooling all compose cleanly on top.
+
+## Get started
+
+```bash
+curl -X POST https://api.tsfm.ai/v1/forecast \
+  -H "Authorization: Bearer $TSFM_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{
+    "model": "amazon/chronos-2",
+    "inputs": [{"target": [10, 12, 11, 13, 14, 15, 14, 16, 18, 17]}],
+    "parameters": {"prediction_length": 24, "quantiles": [0.1, 0.5, 0.9]}
+  }'
+```
+
+```python
+from tsfm import Tsfm
+
+client = Tsfm()
+forecast = client.forecast(
+    model="amazon/chronos-2",
+    inputs=[{"target": [10, 12, 11, 13, 14, 15, 14, 16, 18, 17]}],
+    parameters={"prediction_length": 24, "quantiles": [0.1, 0.5, 0.9]},
+)
+print(forecast.predictions[0].mean)
+```
+
+## Benchmarks
+
+We publish continuously-updated leaderboards scoring every hosted model across **GIFT-Eval** and the **Impermanent** long-horizon benchmark suite:
+
+- [GIFT-Eval leaderboard](https://tsfm.ai/benchmarks/gift-eval)
+- [Impermanent leaderboard](https://tsfm.ai/benchmarks/impermanent)
+
+## Learn more
+
+- [Blog](https://tsfm.ai/blog) — model deep-dives, benchmark analyses, architectural notes.
+- [Use cases](https://tsfm.ai/use-cases) — demand planning, energy load, telemetry, financial time series.
+- [Integrations](https://tsfm.ai/integrations) — Python SDK, Node SDK, CLI, MCP tools.
+
+## Contact
+
+- **General:** [hello@tsfm.ai](mailto:hello@tsfm.ai)
+- **Enterprise:** [sales@tsfm.ai](mailto:sales@tsfm.ai)
+- **Website:** [tsfm.ai](https://tsfm.ai)