fa_get_file_content
ChatGPTRetrieves the full or partial contents of a file stored in the FilesAnywhere system. Use this tool to access and return file data for reading, analysis, summarization, or structured processing. It supports a range of file types, including text documents, spreadsheets, PDFs, and structured data formats When to use: • When a user requests to view or summarize a file • When AI needs file content for analysis, extraction, or transformation • When processing workflows require reading stored documents • When previewing or validating uploaded file data Key Behavior: • Enforces user-level and role-based access control • Supports multiple file types (text, PDF-extracted text, CSV, JSON, etc.) • May return partial content based on parameters (e.g., start line, byte range) • Optimized for both small and large file retrieval scenarios • Does not modify file content—read-only operation Example Use Cases: • “Show me the contents of Q3_report.pdf” • “Extract text from invoice_1042.docx” • “Read the first 200 lines of log.txt for debugging” • “Load CSV data for processing in workflow automation” Example 1: input: Show me the file abc.pdf { file_path: "myfiles/abc.pdf" } output: { content: "...", file_path: "myfiles/abc.pdf" } Example 2: input: Show me the file image.png { file_path: "myimages/image.png" } output: { url: "...", mime_type: "...", file_path: "myimages/image.png" }
fa_get_user_profile
ChatGPTRetrieves the authenticated user’s profile information, including account details such as name, email address, storage usage, and client ID. Use this tool when user-specific account data is required to support identity verification, account display, or retrieval of profile-related information. When to Use This Tool: • When personalizing a response (e.g., addressing the user by name or role) • When retrieving user preferences (language, timezone, UI settings) • When performing account-aware operations (permissions, roles, subscription tier) • When pre-populating forms or workflows with known user data • When confirming identity-related context in multi-step processes Returns May Include (depending on permissions): • User ID (non-sensitive internal identifier) • Display name • Email address (if permitted) • Client ID • Localization preferences (language, timezone, region) • Plan Type Security & Privacy Notes: • Must only return data for the authenticated user context • Must not expose sensitive authentication credentials (passwords, tokens, secrets) • Should respect data minimization rules (only return fields necessary for the request) • Must comply with user consent and organizational privacy policies Example: What is my current logged in user? input: {} output: { user_name:"...", display_name: "...", user_id:..., email:"...", client_id:... }
fa_search_content
ChatGPTSearch FilesAnywhere retrieves files, folders, documents, metadata, or text snippets from stored data based on user-supplied queries. It can be used to locate content using criteria such as keywords, folder path, date range, file type, size, or tags. The results return matching items along with associated metadata, including file name, location path, type, and size. Use this tool when the user requests to: • Find files containing specific text or keywords • Locate documents related to a topic, project, or user • Search within stored file content (not just file names) • Filter results by metadata such as date, file type, or folder location • Retrieve relevant documents for summarization, preview, or further processing The tool returns a ranked list of matching files with contextual relevance indicators, allowing downstream tools to open, preview, or summarize selected results. Key Capabilities: • Full-text search across supported file formats • Metadata-based filtering (file type, owner, date, folder) • Relevance-ranked results • Supports enterprise-scale indexed content • Works across shared and private user drives (respecting permissions) When NOT to use: • When the user requests a specific known file ID (use direct file retrieval instead) • When browsing folder structures without keyword search (use folder listing tools instead) Example: Find pdf files in my account input: { file_type:"pdf_files", include_files: true, include_folders: false } output: { instructions_for_client: ":contains special instructions for the client describing the recommended output style", total_matching_results: 1 /total number of matched pdf files/ files: [ { file_name: "abc.pdf", file_path: "myfiles/abc.pdf", file_directory_path: "myfiles/", file_type: "pdf_files", mime_type: "application/pdf", file_size: "5.2MB", modified_on: "2026-04-02 04:14:00 PM", file_url: "...", is_group_shared: false } ] }
fa_summarize_file
ChatGPTProvides a concise, structured summary of a file’s content. Use this tool when a user uploads or references a document and requests an overview, key points, executive summary, or condensed version. The tool extracts and organizes the most relevant information while preserving meaning and context. It can be used with common business and productivity file types such as PDF, DOCX, TXT, CSV, and other text-based formats. For structured data files like CSV, it summarizes patterns, trends, and key fields rather than reproducing raw data. When to use this tool: • The user uploads or references a file and asks for a summary, overview, or key points • The user wants a simplified explanation of a document’s contents • The user requests “what this file says,” “summarize this document,” or similar phrasing • The user needs insights, highlights, or bullet-point breakdowns of a file • The file content is too long or complex to reasonably present in full Input Requirements: • File identifier or file path from FilesAnywhere • Output: • A structured natural-language summary of the file • Key points and important findings (when applicable) • Optional categorized breakdown depending on summary type Behavior Guidelines: • Do not return raw file contents unless explicitly requested • Preserve accuracy and meaning while condensing information • Prioritize clarity, relevance, and actionable insights • Maintain neutral and professional tone unless otherwise specified Example 1: input: Show me the file abc.pdf { file_path: "myfiles/abc.pdf" } output: { chunks: ["...", "...", "...",...], file_path: "myfiles/abc.pdf", file_name: "abc.pdf", file_url: "..." } Example 2: input: Show me the file image.png { file_path: "myimages/image.png" } output: { url: "...", mime_type: "..." }
