Getting started

1. Open ULTRASEARCH and press the folder button in the top bar.
2. Add one or more root folders in Supervised repositories.
3. Give every repository a unique ID and choose its identifying colour.
4. Press Close & index and wait for indexing to finish.
5. Enter a term, symbol, phrase or structured query in Search.
6. Open results in the Code Viewer or continue with Connections and Insights for structural analysis.

After a successful persistent index has been saved, ULTRASEARCH can restore it in later sessions without rebuilding everything. If the source tree changes, use Update data to scan the configured repositories again.

Indexing & repositories

Supervised repositories let you analyse several independent root folders as one combined workspace. Each root has its own ID and colour, which are reused throughout Search, Connections, Code Viewer, Insights and Git.

• Repository IDs must be unique.
• Supervised roots cannot be the same folder, overlap, or be nested inside one another.
• Update data reads all configured roots again and rebuilds the combined index.
• Project and solution membership is detected where supported and can be shown as context chips.
• The index is stored locally in the browser so it can be restored after reload or reopening the application.

The indexer recognises common source and text formats including C#, VB.NET, JavaScript, TypeScript, Python, Java-family and C-like languages, Go, Rust, PHP, SQL, XML, HTML, Razor, XAML and configuration files. Symbol and relationship extraction is heuristic and depends on the language and source structure.

Search

Search is the main entry point for locating text, files, declarations, symbols and indexed relationships.

Search syntax

Quick operators can be typed directly or inserted with the buttons beside the search controls. Operators can be combined in the same query.

The Clear quick-operator button removes the current quick prefix while keeping its value.

Query builder & search tools

Query builder

Query builder creates advanced filters without requiring you to memorise syntax. Select a filter, enter its value, optionally exclude it, and press Add filter. Repeat this as many times as needed. The generated query remains editable by hand.

Add to current query appends the generated query to Search. Replace query replaces the current Search text.

Search tools

• Matching options include Regular expression, Whole word and Exact identifier.
• Scope can cover the whole codebase or a current namespace, with temporary exclusions.
• Results can be grouped by file, folder, extension or match type and displayed at normal, compact or detailed density.
• Project presets are available for .NET, Web, PHP, C / C++, Scripts and Configuration.
• Search history can rerun previous queries, and Compare searches compares two recorded searches.

Search results

Each result block identifies the source file and labels detected occurrences such as definitions, instantiations, calls, references, imports, comments and strings. Search highlights follow the active theme and can use per-type colours.

• Open a result to inspect the file in the Code Viewer.
• Use Track file or Track symbol to keep useful items in Tracking.
• Use Compare on two result files to open the full-screen file comparator.
• Repository, solution and project chips show where a result belongs; Details appears when multiple project or solution memberships are detected.
• The extension selector is populated from extensions that are actually present in the current results.

Tracking

Tracking is a persistent workspace for searches, files, symbols and personal notes that you want to keep available.

• Add notes or terms manually, or track searches, files and symbols from other sections.
• Assign multiple tags and click a tag to filter tracked items.
• Suggest tags analyses tracked items and proposes structural tags such as architecture, dependencies, complexity, risk, tests, language and symbol kind.
• Create custom tags, remove individual tags, or delete all tags without deleting the tracked items themselves.
• Only tags with items hides unused tags, while Show descriptions explains tag categories.
• Export as JSON creates a portable tracking backup; Import restores compatible tracking data.
• Clear all removes the tracked items after confirmation.

Connections

Connections visualises relationships found in the indexed codebase. The left structure tree is used to browse namespaces, classes and members, while the graph shows the currently selected architectural view.

Select a graph node to see its importance, isolation, fragility, complexity, namespace, source path and visible relationships. When source is available, the node can be opened directly in the Code Viewer.

Code Viewer

The Code Viewer opens source without leaving the current section. It provides syntax highlighting, file tabs, symbol navigation and relationship tools while preserving the original source text.

• Use Back and Forward to move through viewer navigation history, and Prev / Next to move between search matches.
• Go to definition attempts to locate the indexed declaration for the selected identifier.
• Find references, Find callers, Find callees and Inheritance run relationship searches from the selected symbol.
• Double-click an identifier to attempt direct navigation to its definition. When several definitions are plausible, ULTRASEARCH can ask you to choose.
• Symbols that have no indexed declaration can be identified as external rather than opening an unrelated text match.
• The permanent Outline can be filtered and used to jump to declarations. Additional symbol groups may include namespaces, types, methods, properties, events and using/import information.
• Split view opens a second file beside the current file.
• Compare opens the full-screen comparator. Show differences highlights changed lines and inline edits; Sync scroll keeps both files aligned.
• Copy provides source-copy options and Connections opens the relevant graph context.
• The viewer width and outline width are resizable. Minimap can be shown or hidden, and the viewer can be expanded to fill the window.
• Middle-click a Code Viewer file tab to close that tab.

Insights

Insights builds higher-level project intelligence from the current index. Press Analyse project to calculate structural metrics before using the detailed panels.

• Summary cards show indexed files, average structural risk, dependency cycles and dead-code candidates.
• Impact & health ranks hotspots using importance, fragility, blast radius and combined risk.
• Watch queries & regression stores selected searches and can refresh their counts to reveal changes.
• Architecture rules & zones lets you map path fragments to named zones and define forbidden dependency directions between zones.
• Snapshots preserve dependency structure so Compare latest can show added and removed dependencies.
• Recent workspaces keeps locally recorded project summaries.
• Structural intelligence detects modules, structurally similar files and duplicate architecture patterns. Similar files can be opened directly in the file comparator.
• Export report saves an HTML architecture report containing health statistics, hotspots and architecture information.

Git

Git Intelligence scans the supervised repository roots recursively for readable local Git metadata, including nested repositories. Press Analyse Git repositories to run or refresh the scan.

• Each detected repository is associated with its supervised repository chip.
• The current implementation reads local HEAD log information to build a recent repository timeline and activity estimate.
• After a reload, the browser may require folder permission again before Git metadata can be read.
• Browser-only analysis does not run a native Git executable. Packed Git objects therefore do not always expose reliable per-file commit churn.

History

History records useful navigation actions performed during the current and stored application workflow.

• Recorded types include Search, Code view, Graph node and Section.
• Use the Type selector to filter the history.
• Select an entry to reopen the corresponding search, file, node or section.
• Opening an action from History does not create another identical history entry, and consecutive identical actions are not duplicated.
• Individual entries can be removed, or Clear history can remove the full recorded history after confirmation.

Options & stored data

Options controls what is indexed and how locally stored application data is managed.

Themes, languages & zoom

The Theme selector changes the complete interface, including Search, Code Viewer, graphs, dialogs, analysis indicators and this guide. Help cards, headings and text use the same theme variables as the rest of ULTRASEARCH so foreground content remains readable across dark, light, high-contrast, vibrant, colour and multicolour themes.

The Zoom selector scales the application interface from 80% to 150%. The startup loading screen remains at its normal size.

English is the source language and Spanish is fully reviewed locally. Chinese, Czech, Dutch, Finnish, French, German, Hindi, Irish, Italian, Japanese, Korean, Norwegian, Polish, Portuguese, Russian, Swedish and Ukrainian are also supported. Reviewed local strings are used first; missing strings can be completed by the browser's local automatic translation system when available, with English as the fallback.

Keyboard & navigation

Local operation & privacy

• Project files are read and analysed locally by the application.
• Normal indexing, searching, Code Viewer navigation, Connections and Insights do not upload project source files.
• Persistent indexes, Tracking, saved graph views, Insights snapshots, settings and History are stored locally in browser storage.
• Export actions only create files when you explicitly choose to save them.
• Automatic translation for missing interface strings uses the browser's local translation capability when available; otherwise the original English text remains visible.
• Indexing and analysis never modify the source files on disk.

Troubleshooting