File Tools

The File provider lets you work with files at every stage of a workflow: parse uploaded or URL-hosted documents into text, download and parse remote files, create new workspace files from generated content, or append text to existing files. Use these tools whenever a workflow needs to ingest, produce, or update file-based data.

Overview

Operations

Configuration

file_parser

Parses one or more files. Accepts a file path string, an array of paths, or an external URL. Supports PDF, CSV, DOCX, Word/Excel, CAD formats (DWG, DXF, STEP, IGES), images, plain text, and Markdown. File type is auto-detected when not supplied.

Notes:

• When the block input method is set to Upload, filePath is derived from the uploaded file objects automatically — you do not need to set it manually.
• The fileType parameter is hidden in the UI and is rarely needed; the parser identifies file types automatically.

file_write

Creates a new file in the workspace Files store from a text string. The file becomes visible to every workflow in the workspace.

file_append

Appends text to the end of an existing workspace file. If the named file does not exist it is created automatically.

Outputs

file_parser

• files (array) — Array of parsed file objects. Each object contains:
  • content (string) — Extracted text content of the file.
  • fileType (string) — Detected file type (e.g. pdf, csv).
  • size (number) — File size in bytes.
  • name (string) — Original file name.
  • binary (boolean) — Whether the file was treated as binary.
  • metadata (object, optional) — Additional file-specific metadata (e.g. page count for PDFs).
• combinedContent (string) — All file contents merged into a single text string, separated by dividers. Convenient when you want to pass the full content of multiple files to an LLM block.
• file1, file2, … (object) — Named shortcuts to individual file results in multi-file parses (e.g. {{file_block.file1.content}}).

file_write

• id (string) — Workspace file ID assigned to the newly created file.
• name (string) — Final file name after any deduplication applied by the platform.
• size (number) — File size in bytes.
• type (string) — MIME type of the created file.
• url (string) — URL to access or download the file.

file_append

• id (string) — Workspace file ID of the file that was appended to.
• name (string) — File name.
• size (number) — New file size in bytes after the append.
• type (string) — MIME type of the file.
• url (string) — URL to access or download the file.

YAML Example

file_1:
  type: file
  name: "File"
  inputs:
    operation: "file_parser"
    filePath: "https://example.com/report.pdf"
  connections:
    outgoing:
      - target: next-block-id

file_2:
  type: file
  name: "File"
  inputs:
    operation: "file_write"
    fileName: "summary.md"
    content: "{{agent_1.content}}"
  connections:
    outgoing:
      - target: next-block-id

Tips

• No API key required. The File tools call internal Zelaxy APIs (/api/files/parse, /api/files/write, /api/files/append). Auth is handled by the workspace session. workspaceId is injected automatically from the execution context — you never need to supply it.
• Batch parsing. Pass an array of URLs or upload multiple files at once. The combinedContent output merges all text with clear dividers, while individual files are accessible via {{file_block.files[0].content}} or the named shortcuts {{file_block.file1.content}}, {{file_block.file2.content}}, etc.
• Write then read in the same workflow. file_write returns a url you can pass directly into a downstream file_parser block to verify or further process the file you just created.