Content Explorer
Overview
Content Explorer helps you quickly understand large sets of documents without reviewing files one by one. It treats a folder or entire Workspace as a single corpus and generates an AI-written Corpus Summary of the indexed content. It also extracts entities across the corpus so you can review recurring people, organizations, locations, dates, and other identifiers, then explore those relationships in an Entity Graph. You can also investigate specific claims in the summary using Explain / Find Sources. Inputs are indexed documents from your selected folder or Workspace. Outputs include Corpus Summaries, entity lists, entity visualizations, Entity Graphs, source-tracing results, and exported RTF summaries. AI-generated summaries and source tracing improve speed and coverage, but results should always be validated against original documents.
When to Use This Application
- You need a high-level summary of many related documents before reviewing them individually
- You want to identify recurring people, organizations, places, or structured identifiers across a corpus
- You need clickable references that connect summaries back to source material
- You want to visualize relationships between entities across multiple documents
- You need to clean noisy entity extraction results before deeper analysis
- You want to investigate specific claims without manually searching documents
Before You Begin
- Before starting, ensure your documents are available in Octostar and have completed indexing. If documents are still indexing, they may not appear in summaries or entity analysis.
- Confirm you have access to the relevant folder or Workspace, and verify that AI features are enabled if you plan to use summaries, graph summaries, or source tracing.
- If needed, open Corpus Options before starting to adjust summary language, model selection, response length, or concurrency settings. You can also expand your corpus during a session by dragging additional files or folders into the app.
Step-by-Step Walkthrough
Step 1 — Set corpus scope
Launch Content Explorer from a folder or Workspace.
After loading, the header displays the corpus scope along with file count, indexed page count, file size, and oldest/newest file indicators so you can confirm you are analyzing the correct material.
Step 2 — Build the corpus
After files are added, Content Explorer automatically begins preparing the corpus by scanning files, identifying indexed content, and aggregating entities. Once preparation finishes, the app automatically generates the first summary using your saved settings. You may see progress indicators, status messages, or indexing warnings during this process. If some files are not indexed, they may be excluded from summaries and entity analysis until indexing completes.
Step 3 — Configure summary settings (optional)
You can adjust summary length depending on how much detail you need, ranging from shorter summaries to multi-page outputs.
The Summary Purpose menu allows you to choose between:
| Standard Fast | Quicker balanced run (skips some deeper evaluation passes) |
|---|---|
| Standard | Fuller evaluation across documents and chunks |
| Focus on persons, location, crimes or risks | Narrows the briefing. Edit the prompt text to name specific subjects when needed. |
| Custom | Your own purpose prompt from scratch |
The Corpus Options menu allows you to change:
- summary language,
- chat model,
- maximum response length,
- maximum concurrent AI requests. Changing these settings automatically starts a new summary once the corpus is ready.
Use Regenerate when you want to rerun the summary using the same settings.
Step 4 — Review the Corpus summary and validate references
The summary appears as a structured narrative with numbered reference chips (for example [1], [2]) that point to sources. Hover or click a chip to see which document, where in the document, a short excerpt, and Open document.
If some files were not referenced in the summary, you can expand the notification panel to review them manually.
Step 5 — Use Explain / Find Sources (Optional)
Highlight a phrase or sentence in the summary and select Explain / Find Sources to investigate a specific claim. The tool generates a targeted explanation and provides supporting document references so you can validate important findings more quickly.
Step 6 — Review extracted entities
The Entity Panel on the right side, allows you to review extracted entities by category, including people, organizations, locations, dates, phone numbers, email addresses, and URLs. You can search entities, sort them by frequency or document coverage, and switch between table and word cloud views depending on how you want to analyze the results.
Step 6 — Clean entity results (Optional)
Use the Entity Cleanup tool (lighting icon in the entity panel) when duplicate names, formatting inconsistencies, or extraction noise make results difficult to interpret. You can merge similar entities, remove irrelevant entries, and improve the quality of both entity counts and graph analysis.
Step 8 — Explore the Entity graph
The Entity Graph helps you visualize relationships across documents.
Use minimum pages to hide entities that appear only once or twice if the graph is too busy.
You can switch between entity-to-entity and entity-to-document views, filter low-frequency entities, hide categories, search for specific entities, and navigate the graph through pan and zoom controls.
Selecting a node opens additional details, including related documents and short AI-summaries that explain how the entity appears across the corpus.
Step 9 — Use debugging tools (Optional)
The settings menu includes debugging tools such as console logs, chat logs, corpus inspection tools, and API call logs. These tools are intended for troubleshooting and advanced workflows rather than everyday use.
Understanding the Output
The primary output is the Corpus Summary, which provides a narrative overview with citations and document links that help you move from summary to evidence. Entity tables and word clouds provide a quantitative view of recurring people, organizations, and identifiers across the corpus. The Entity Graph adds a visual layer for exploring relationships, while Explain / Find Sources helps validate individual claims.
Saving and Exporting Results
Use Copy to place summary text on your clipboard. Use Save to Workspace to export the summary as an RTF file and store it in your Workspace for future reference. The summary toolbar may also display word and character counts for the current output.
Tips for Best Results
- Use focused document sets when possible
- Increase summary length when deeper analysis is needed
- Fix indexing gaps before relying on results
- Use custom prompts for specialized investigations
- Clean entities before analyzing graph relationships
- Reduce concurrent AI requests if processing becomes unstable
Known Limitations
| Limitation | Explanation |
|---|---|
| Indexed content only | Files not indexed will be skipped; entity views and summaries depend on the same index. |
| AI text is not evidence | Overviews, graph node blurbs, and Find Sources narratives can be incomplete — always verify in source files. |
| Automated entity noise | Extracted names include errors and duplicates; counts and graphs improve after manual cleanup. |
| Custom purpose required | Custom mode needs non-empty instructions or the app will not generate a summary. |
| Heavy corpora and services | Very large scopes or busy backends can slow preparation or trigger timeout messages. |
| Cleanup timing | Merge suggestions may load asynchronously; the wizard waits until data is ready. |