---
title: "Understand Dataset Indexing and Time Estimates"
description: "Index a dataset for semantic search in Noema, read the remaining-time estimate, and recover from indexing problems."
version: "Noema 4.1+"
platforms: ["iPhone", "iPad", "Mac", "Vision Pro"]
reviewed: "October 3, 2026"
canonical: "https://noemaai.com/docs/dataset-indexing"
---

# Understand Dataset Indexing and Time Estimates

How embedding works, why the time estimate can start as …, and what to check when indexing stalls.

Indexing prepares your documents for semantic search. A chat model generates answers; an embedding model creates the searchable representations used to retrieve relevant passages.

## Index a dataset

1. Import or download your material.
2. Allow extraction or transcription to finish.
3. Install the required embedding model.
4. Open the dataset’s detail page.
5. Choose **Start Embedding Process**, or **Rebuild Dataset Index** when required.
6. Wait until the dataset reports **Ready**.

Depending on the material, progress can include extraction, compression, and embedding.

## Read the remaining-time estimate

Noema 4.1 estimates remaining embedding time from completed work.

At the beginning, the estimate may display **…** while Noema collects enough measurements. This is expected: preparation and warmup do not provide a reliable measure of indexing speed.

The estimate can change as the remaining material or processing speed changes. After recovery, Noema may temporarily withhold the estimate while it measures throughput again.

A missing estimate alone does not mean indexing has stopped. Check the processing stage, completed progress, and any error message.

## Recover an indexing problem

Open the dataset’s detail page and review its status.

- Confirm that the embedding model is available.
- Resolve any storage or extraction error.
- Start or rebuild the index when the page offers that action.
- Wait for **Ready** before relying on retrieval or sharing the dataset.

For a dataset hosted by another Mac, perform index maintenance on that Mac.

## Related documentation
- [Datasets & RAG](https://noemaai.com/docs/dataset-integration)
- [Search Datasets Shared by Your Mac](https://noemaai.com/docs/mac-shared-datasets)
- [Knowledge Packs](https://noemaai.com/docs/knowledge-packs)