Skip to content
Nappai Documentation

Gemini File Search Store

This component acts as a smart bridge between your files and Google’s AI. It automatically organizes, indexes, and searches through your document collections without requiring any technical setup. When you ask a question, it finds the most relevant parts of your documents and returns clear answers with automatic citations, making it easy to verify where the information came from.

How it Works

This component connects directly to Google’s File Search API to manage your document collections. Instead of asking you to set up databases or configure complex indexing rules, it handles everything in the background.

When you add files, the component automatically splits them into readable sections, converts them into a format AI understands, and stores them securely in Google’s cloud. When you run a search, it quickly compares your question to your stored documents, finds the best matches, and returns the exact text snippets along with automatic references to the original files. This means you get accurate, traceable answers without managing servers or vector databases yourself.

Connection & Credentials

This component requires configuring a credential in the Nappai panel before interacting with the external service:

  1. Go to the Credentials section in your Nappai panel.
  2. Create a new credential of the type Google Gemini and fill in the required fields (API Keys, tokens, etc.).
  3. In your workflow, select the saved credential in the Credential input field of this node.

Inputs

Configuration for this component is primarily handled through the Credential field. All other technical parameters, such as document paths, query text, and pagination settings, are managed automatically by the platform’s base configuration. This streamlined approach allows you to focus on selecting your data source and running searches without dealing with complex setup forms.

Outputs

The component returns structured data that can be easily connected to other workflow steps. You will receive confirmation that your documents were successfully saved and indexed, a list of your active document collections, and detailed search results.

Search results contain the exact text snippets that matched your query, relevance scores to show how closely they align with your question, and automatic citations linking back to the original files. This makes it simple to verify information or feed the results into other tools for reporting or AI generation.

Output Data Example (JSON)json

{ “corpus_status”: “indexed”, “store_id”: “nappai_corp_123”, “search_results”: [ { “file_id”: “doc_abc”, “snippet”: “The system automatically processes uploads and splits them into manageable sections for faster searching.”, “similarity_score”: 0.92, “citation”: “Project_Manual_v1.pdf, page 15” }, { “file_id”: “doc_def”, “snippet”: “Users can reset their passwords through the settings dashboard or by contacting support.”, “similarity_score”: 0.87, “citation”: “User_Guide.pdf, page 8” } ], “metadata”: { “total_documents”: 5, “last_updated”: “2024-01-15T10:30:00Z” } }

Connectivity

This component is typically placed in the middle of a data automation pipeline.

  • Input Connections: It usually receives data from File Upload, Document Parser, or Data Extractor components that prepare your files for indexing.
  • Output Connections: Its search results and document status outputs are commonly connected to LLM Generators (to draft final answers), Data Visualization tools (to display relevant snippets), or Data Routing components (to trigger actions based on search findings). This flow makes logical sense because the component first indexes your documents, retrieves the most relevant text, and passes that verified information forward for final processing or presentation.

Usage Example

Scenario: Automating Customer Support Q&A

  1. Upload product manuals and FAQ PDFs using a File Upload component.
  2. Connect these files to the Gemini File Search Store to automatically index them into a “Support Knowledge Base”.
  3. Later, run a search using a question like "How do I reset my password?".
  4. The component returns exact excerpts from your manuals along with citations.
  5. Connect these excerpts to a Text Display component for quick reference, or to an AI Generator to create polished, citation-backed responses for customers.

Tips and Best Practices

  • Keep your document collections focused on specific topics to improve search accuracy.
  • Use clear, descriptive file names before uploading, as they often help the system organize information more effectively.
  • Review the similarity scores in the output to ensure the returned snippets are highly relevant to your query before using them in reports.
  • Leverage the automatic citations to quickly verify information without manually cross-referencing documents.

Security Considerations

  • Your Google API Key is securely stored in Nappai’s credential manager and is never exposed in workflow logs or data exports.
  • All document indexing and search operations occur within Google’s secure infrastructure, ensuring your data remains protected according to industry standards.
  • Regularly review and rotate your API keys in your Google Cloud console if you suspect any unauthorized access.
  • Ensure that uploaded documents comply with your organization’s data privacy and content policies before indexing them into the system.