Mastering Dhito's Search Filters: Finding Files Faster

Semantic search fundamentally changes how we find information. Instead of relying on exact keyword matches, Dhito understands the *meaning* of your query. However, even with an intelligent vector space, sometimes you have specific constraints in mind. You might remember that the document you need is specifically a PDF, or that you saved it just last week.

To bridge the gap between open-ended semantic exploration and precise retrieval, Dhito offers powerful Search Filters built right into the interface.

Why Filters Matter in an AI Search Engine

When you search for "Q3 revenue projections," a pure semantic search might return: * A PDF board pack containing the actual figures. * A recorded Zoom meeting where the projections were discussed. * A Slack screenshot where someone mentioned the numbers.

While it is useful that Dhito can surface all of these across different modalities, you might *only* want to see the document. This is where filters come in.

Using the Filter Tabs

Dhito makes filtering intuitive by keeping it visual. You do not need to memorize complex query languages or boolean operators. Instead, just look below the search bar.

1. File Type Filters

Directly below the search input, you will find tabs for different file modalities. Clicking these tabs instantly narrows your search results:

• Documents: Restricts results to written formats — PDF, Word (DOC and DOCX), RTF, ODT, EPUB, plain text, Markdown, RST, CSV and TSV, plus HTML, JSON, YAML and around 33 code formats. Perfect for finding contracts, essays or reports. Worth knowing before you rely on it: Excel, PowerPoint, Pages, Keynote and Numbers files are not indexed at all, so they will never appear under this tab.
• Images: Narrows the search to JPEG, PNG, HEIC, HEIF, WebP, PSD, TIFF, BMP, ICO, AVIF and ten camera RAW formats, including Photoshop files and camera RAW straight off a memory card. Since Dhito uses the Florence-2 vision model, it will search the *visual content* and *text inside* these images.
• Audio & Video: Limits results to media — 17 video containers and 14 audio containers, from MP4 and MOV to MKV, WebM, MP3, WAV, M4A and FLAC. Because Dhito transcribes these locally with Whisper and indexes the transcript at segment level, your semantic query matches the spoken words *and* lands on the minute where they were said.

By selecting one of these tabs, you tell Dhito: "Find the semantic meaning of my query, but *only* within this specific bucket of files."

2. Date Filters

Sometimes the context of *when* a file was created or modified is just as important as what is inside it. Next to the file type filters, you can access date ranges to refine your search chronologically:

• Past Week: Ideal for finding that recent draft or a meeting recording from a few days ago.
• Past Month: Great for quarterly reviews or finding recent project assets.
• Past Year: Helps filter out legacy documents when you only want relatively recent context.

Combining Filters for Maximum Precision

The true power of Dhito's interface is combining these filters with semantic search.

For example, you can search: "design feedback for the landing page" Then, click the Images tab, and select Past Week.

Instantly, Dhito will filter out all text documents and older assets, surfacing only recent screenshots or design mockups related to your landing page feedback.

Conclusion

While Dhito's local AI models do the heavy lifting of understanding your intent, the filter tabs put the final control back in your hands. By combining the intelligence of semantic search with the precision of file type and date filters, you can find the exact file you need faster than ever before.