> For the complete documentation index, see [llms.txt](https://raic-labs.gitbook.io/raic-fed/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://raic-labs.gitbook.io/raic-fed/analyst-guide/03c-running-your-first-search.md).

# Running your first search

With data ingested, the Analyst is ready to run a search. RAIC Fed searches an entire imagery collection using one or more reference images, with no labels or pre-trained classes required. Results are clustered by semantic similarity in a shared vector space.

<figure><img src="/files/7mRb7AC9yitRQOOwjqj2" alt=""><figcaption></figcaption></figure>

***

## The New Search panel

Select **New Search** from the top navigation to open the search configuration panel. The panel has four sections, each collapsible:

* **Temporal Extent** — the date range to search across.
* **Spatial Extent** — the geographic area to search within.
* **Detection Filters** — the Detector to apply to the search.
* **Reference Image** — the seed image or taxonomy category that tells RAIC Fed what to look for.

Work through each section top to bottom before running the search.

***

## Step 1: Set the temporal extent

The Temporal Extent defines the time window RAIC Fed will search across.

1. In the **Temporal Extent** section, select a preset time range: **Last Year**, **Last 6 Months**, **Last Month**, **Last Week**, or **Today**.
2. To set a custom range, enter a **Start Date** and **End Date** directly in the date fields.

> **Analyst** Start with a broader time range on the first search. A narrower window can always be applied once the Analyst has a sense of how results are distributed across the dataset.

***

## Step 2: Set the spatial extent

The Spatial Extent defines the geographic area to search within. By default, the search covers the full dataset. To constrain it to a specific area:

1. In the **Spatial Extent** section, activate the polygon tool by selecting the **Draw polygon** icon on the right side of the map.
2. Draw one or more polygons directly on the map to define the areas of interest. Alternatively, paste GeoJSON or select **Import** to upload a file from your computer.
3. The panel will display the number of polygons drawn and the total combined area in KM².
4. To remove a polygon and start over, select the clear icon next to the area reading.

<figure><img src="/files/wVJspUzuXiVGmbLSDg7Q" alt=""><figcaption></figcaption></figure>

> **Analyst** For a first search, leaving the spatial extent unconstrained is recommended. Broad-area search is one of RAIC Fed's core strengths; narrowing too early can cause the Analyst to miss relevant detections outside the expected area.

***

## Step 3: Select a Detector

The Detection Filters section controls which Detector RAIC Fed applies to the search.

1. In the **Detection Filters** section, select the **Detector** dropdown.
2. Choose the Detector that matches the data that was ingested and the mission objective. Only Detectors compatible with the ingested data will be active in the dropdown. The Detector dropdown will provide the number of scenes for a given detector. This helps you avoid picking a detector that has no underlying data in your current area/time range.

> **Note** If the expected Detector is not active, confirm with the Admin that the data was ingested with that Detector enabled.

***

## Step 4: Select reference image(s)

The reference image is used for RAIC Fed search. It can be an images uploaded by the Analyst or a set of images from an existing taxonomy category.

**To upload an image:**

1. In the **Reference Image** section, select the **Upload** tab.
2. Drag and drop an image into the upload area, or select it to browse.
3. The image will appear in **My images** once uploaded and is available for future searches.
4. Select one or more images from **My images** to use as the reference.

**To use images from existing taxonomy categories:**

1. Select the **Categories** tab.
2. Browse and select a category. All images in that category will be used as the reference set.

**To select a reference image directly from ingested imagery:**

The Analyst can also find an object within ingested data and use it as the reference image without leaving RAIC Fed.

1. From the left navigation menu, select the **list icon** to open the ingested files list.
2. Locate the TIFF file containing the object of interest. Use the **Filter by file name** field to search if needed.
3. Select the **globe icon** next to the file to display it on the map.
4. Select the **magnifying glass icon** on the right side of the map. RAIC Fed will zoom to the supported zoom level automatically.
5. Choose a selection option from the toolbar that appears:
   * **128x128**, **256x256**, or **4096x4096** — a fixed-size box for Tile Mode search.
   * **Free-form** — draw a custom bounding box around the object of interest.
6. Draw a bounding box around the object of interest or move the fixed-box size for Tile Mode.
7. In the dialog that appears, select **Save to My Images**.
8. The image is now available under **My Images** in the **Reference Image** section of the New Search panel.

{% hint style="info" %}
Selecting one of the fixed tile sizes (128x128, 256x256, 4096x4096) starts a selection for a Tile Mode search instead of a standard object detection search. See **Searching with Tile Mode** for that workflow, including which Detector to pair with each size.&#x20;
{% endhint %}

<figure><img src="/files/yP594bTheQwAEeXLowYp" alt=""><figcaption></figcaption></figure>

<figure><img src="/files/4mjXBJzu1cw2FZhv0pXC" alt=""><figcaption></figcaption></figure>

> **Analyst** The reference image does not need to be a perfect match. A clear image of the object type from any source is sufficient. The closer the visual characteristics match what is in the ingested data, the stronger the initial results will be.

***

## Running the search

Once all four sections are configured:

1. Select **Run Search** to submit.
2. Search results in a context map will appear. A single reference image typically returns results within seconds; searches using a larger category set may take longer depending on the size of the reference set and dataset.

<figure><img src="/files/fly9G6EhJ82VS3cCUwoU" alt=""><figcaption></figcaption></figure>

{% embed url="<https://scribehow.com/embed/How_to_Perform_an_Image-Based_Search_on_RAIC_AI__ik2hA3Z4TAu2Wl4abI7Biw>" %}

***

## Search history

Select **Search History** from the top navigation to view previous searches. This panel provides a lightweight way to jump back into prior results without reconfiguring a search from scratch.

Searches are organized into two groups:

* **Pinned** — searches the Analyst has chosen to keep. Pinned searches appear at the top and persist until unpinned. There is no limit to the number of pinned searches.
* **Recent** — the last 25 searches, listed by date. Older searches beyond 25 are not shown.

Each search in the list displays the reference image or category name, date range, scene count, and Detector used. Selecting anywhere on a search reopens its context map or right-click any entry to open it in a new tab, so you can compare results side-by-side without losing your place.

### Renaming a search

By default, searches are named by date and time. To give a search a more descriptive name:

1. Select the menu icon on the search entry.
2. Select **Rename** and enter a new title.

### Pinning a search

1. Select the menu icon on the search entry.
2. Select **Pin** to move it to the Pinned section at the top of the list.

### Deleting a search

1. Select the menu icon on the search entry.
2. Select **Delete** to remove it from the history.

> **Note** Search history is personal to each Analyst and is not shared across the workspace.

<figure><img src="/files/kERBu6P9qadpu0aMCURC" alt=""><figcaption></figcaption></figure>

***

## What's next

Results are clustered by semantic similarity, making it possible to analyze large volumes of detections quickly. Continue to \[Reading your search results →] to learn how to interpret, organize, and act on what RAIC Fed surfaces.

**Compliance Note** This documentation does not constitute legal or regulatory guidance. Organizations operating in classified or regulated environments should consult their compliance and security teams before deploying or using RAIC Fed. Specific data handling, classification, and audit requirements will vary by environment and mission.
