Skip to content

System Operations

System operations are the building blocks of Raikoo workflows. They provide programmatic operations for AI interactions, file manipulation, format conversion, web interaction, code execution, and more. This guide documents all available operations organized by category as they appear in the workflow builder.

Operations are deterministic and perform specific functions with defined inputs and outputs. They can be combined with AI operations to create powerful, end-to-end automation workflows.

AI / LLM

Agent

The primary AI operation for Raikoo workflows. Executes an agent with configured system prompts, tools, and models to accomplish complex tasks through iterative reasoning and tool use.

Configuration:

  • Request: The user message or task for the agent to accomplish
  • Parameters:
    • System Prompts: Instructions that guide the agent's behavior and expertise
    • Tools: Available tools the agent can use (e.g., file operations, web search, custom tools)
    • Model: The AI model to use (e.g., Claude Opus 4.6, Claude Sonnet 4.5)
    • Max Tool Roundtrips: Maximum number of tool use iterations (prevents infinite loops)
    • Context Files: Additional files to include in the agent's context
  • Output: The agent's final response and any workspace artifacts created

Example Use Case: Building an automated research assistant that can search the web, analyze documents, write reports, and iterate on the content based on quality criteria.

Create Image

Generates images from text prompts using AI image generation models.

Configuration:

  • Request: The text prompt describing the image to generate
  • Parameters:
    • Model Selector: The AI image generation model to use
    • Output: Path where the generated image will be saved
  • Output: Generated image file

Example Use Case: Creating visual assets based on textual descriptions for marketing materials, illustrations, or design prototypes.

Google Generative AI: Create Video

Generates video content using Google's generative AI video models.

Configuration:

  • Request: The text prompt or configuration for video generation
  • Parameters:
    • Model: Google video generation model to use
    • Duration: Video length parameters
  • Output: Generated video file

Example Use Case: Creating video content for presentations, marketing campaigns, or educational materials from text descriptions.

Image Object Extractor

Extracts and identifies objects from images using AI vision capabilities.

Configuration:

  • Request: Image file to analyze
  • Parameters:
    • Detection Mode: Type of objects to detect (general, specific categories)
    • Confidence Threshold: Minimum confidence level for detections
  • Output: List of detected objects with bounding boxes and confidence scores

Example Use Case: Analyzing product images to automatically tag inventory, or processing security camera footage to identify specific objects.

Multi-step with Alignment and Tools

Legacy AI operation that runs multi-step conversations with system prompts (alignment) and tool access. Similar to Agent but uses the older configuration model.

Configuration:

  • Request: The user message to process
  • Parameters:
    • Alignment: System prompt to guide the AI's behavior
    • Tools: Available tools for the AI to use
    • Model: AI model selection
    • Max Steps: Maximum number of conversation turns
  • Output: AI response and any generated artifacts

Example Use Case: Running multi-turn AI conversations that need tool access for complex tasks. For new workflows, prefer the Agent operation.

Single-step with Alignment

Simple single-turn AI operation with a system prompt. Sends one request and gets one response without tool use.

Configuration:

  • Request: The user message or content to process
  • Parameters:
    • Alignment: System prompt that defines the AI's role and instructions
    • Model: AI model to use
  • Output: Single AI response

Example Use Case: Simple content transformation tasks like summarization, translation, or text analysis where tool use is not needed.

Conversion

Anything to Markdown

Converts various document formats to Markdown automatically. Supports multiple input formats with automatic format detection.

Configuration:

  • Request: Document content in any supported format
  • Parameters: None required
  • Output: Markdown representation of the document

Example Use Case: Normalizing documents from various sources into a consistent Markdown format for further processing.

HTML to Markdown

Converts HTML content to Markdown format.

Configuration:

  • Request: HTML content to convert
  • Parameters: None required
  • Output: Markdown representation of the HTML

Example Use Case: Converting web content or HTML emails to Markdown for cleaner processing or storage.

HTML to Text

Extracts raw text content from HTML, discarding tags and formatting.

Configuration:

  • Request: The HTML content you wish to convert
  • Parameters:
    • Format (optional): Set to "true" to format the extracted text using Prettier
  • Output: The extracted raw text from the input HTML

Example Use Case: Cleaning web content for processing by AI operations that work better with plain text.

Markdown to Word DOCX

Converts Markdown content to Word DOCX format. Supports template-based generation with placeholder variables.

Configuration:

  • Request: Markdown content to convert
  • Parameters:
    • Template (optional): DOCX template file with placeholders
    • Variables (optional): Values to substitute for {{placeholder}} variables
  • Output: Generated Word DOCX file

Example Use Case: Generating professional reports, contracts, or formal documents that require Word format from Markdown content created by AI operations.

PDF to Markdown

Converts PDF documents to Markdown format.

Configuration:

  • Request: A base64 encoded PDF document
  • Parameters: None required
  • Output: The markdown representation of the PDF content

Example Use Case: Extracting content from PDF documents for analysis or editing in workflows.

Sheet to CSV

Converts Excel XLSX to CSV format.

Configuration:

  • Request: A data URI containing a spreadsheet file (XLSX)
  • Parameters:
    • File (optional): A workspace input reference to the spreadsheet file
    • SheetNameOrIndex (optional): Name or index of the sheet to convert
  • Output: A CSV string representation of the specified sheet

Example Use Case: Converting spreadsheet data to a simpler format for data processing or import into other systems.

Sheet to HTML

Converts Excel XLSX to HTML format.

Configuration:

  • Request: A base64 encoded XLSX spreadsheet data URI
  • Parameters: None required
  • Output: The HTML representation of the input spreadsheet

Example Use Case: Creating web-viewable tables from spreadsheet data.

Sheet to JSON

Converts Excel XLSX to JSON format.

Configuration:

  • Request: A base64 encoded XLSX spreadsheet provided as a data URI
  • Parameters: None required
  • Output: A JSON object representation of the provided spreadsheet

Example Use Case: Converting spreadsheet data to a structured format for easier processing in workflows.

Sheet to Markdown

Converts Excel XLSX to Markdown table format.

Configuration:

  • Request: A base64 encoded data URI representing an XLSX spreadsheet file
  • Parameters: None required
  • Output: A markdown representation of the input spreadsheet

Example Use Case: Converting spreadsheet data to markdown tables for inclusion in documentation or reports.

Word to HTML

Converts Word DOCX to HTML format.

Configuration:

  • Request: A base64 encoded data URI of a Word (.docx) document
  • Parameters: None required
  • Output: The HTML content generated from the Word document

Example Use Case: Converting Word documents into web-ready format for publishing or further processing.

Word to Markdown

Converts Word document (.docx) to Markdown format.

Configuration:

  • Request: A base64 encoded data URI representing a Word (.docx) file
  • Parameters:
    • File (optional): A workspace input reference to a Word document (takes precedence over request body)
  • Output: The markdown representation of the Word document

Example Use Case: Converting existing Word documents into a format that can be easily processed by AI operations.

Web

API Call

Makes HTTP API calls with configurable method, headers, body, and authentication. More structured than HTTP Request with built-in auth support.

Configuration:

  • Request: Request body (for POST, PUT, PATCH methods)
  • Parameters:
    • URL: The API endpoint to call
    • Method: HTTP method (GET, POST, PUT, PATCH, DELETE)
    • Headers: HTTP headers as key-value pairs
    • Authentication: Built-in auth configuration (API key, Bearer token, Basic auth)
    • Query Parameters: URL query parameters
  • Output: API response body

Example Use Case: Integrating with external APIs to fetch data, submit forms, or trigger actions in third-party services.

Performs web searches using the Brave Search API. Returns search results that can be processed by subsequent operations.

Configuration:

  • Request: Search query string
  • Parameters:
    • Count: Number of results to return
    • Search Type: Web, news, images, videos
    • Freshness: Time range for results
  • Output: Structured search results with titles, URLs, snippets, and metadata

Example Use Case: Gathering current information from the web for research, content creation, or data enrichment workflows.

BrowserCat: Load and Select HTML

Uses BrowserCat headless browser service to load a URL (including JavaScript-rendered content) and extract elements using CSS selectors.

Configuration:

  • Request: The URL to load
  • Parameters:
    • Selector: CSS selector for elements to extract
    • Wait For: Selector or time to wait before extraction
    • Execute JavaScript: Optional JavaScript to run before extraction
  • Output: Extracted HTML elements after JavaScript execution

Example Use Case: Scraping content from modern web applications that rely heavily on JavaScript rendering, such as single-page applications.

Get HTML

Retrieves HTML content from a URL.

Configuration:

  • Request: The URL you wish to get the HTML of
  • Parameters:
    • Format (optional): Set to "true" to format the HTML using Prettier
  • Output: The HTML content of the specified URL

Example Use Case: Fetching web content for analysis or transformation in workflows.

HTTP Request

Makes general HTTP requests. Flexible operation for calling any HTTP endpoint.

Configuration:

  • Request: Request body (for methods that support it)
  • Parameters:
    • URL: The endpoint to call
    • Method: HTTP method
    • Headers: HTTP headers
  • Output: Response body

Example Use Case: Making custom HTTP requests to APIs or web services that don't require complex authentication.

Load and Select HTML

Loads HTML from a URL and extracts specific elements using CSS selectors. Similar to BrowserCat version but uses a simpler HTTP fetch.

Configuration:

  • Request: The URL to load
  • Parameters:
    • Selector: CSS selector to extract elements
    • Format (optional): Format the output HTML
  • Output: Array of matching HTML elements

Example Use Case: Extracting specific content from static web pages without needing full browser rendering.

File Operations

Copy Directory

Copies an entire directory and its contents to a new location in the workspace.

Configuration:

  • Request: Source directory path
  • Parameters:
    • Destination: Target directory path
  • Output: Copied directory in workspace

Example Use Case: Duplicating directory structures for parallel processing or creating backups before modifications.

Copy File

Copies a file to a new location in the workspace.

Configuration:

  • Request: Source file path
  • Parameters:
    • Destination: Target file path
  • Output: Copied file in workspace

Example Use Case: Creating file copies for versioning or processing variations of the same file.

Delete Directory

Deletes a directory and all its contents from the workspace.

Configuration:

  • Request: Path to the directory to delete
  • Parameters: None required
  • Output: Success confirmation

Example Use Case: Cleaning up temporary directories or removing processed content from the workspace.

Delete File

Deletes a file from the workspace.

Configuration:

  • Request: The path to the file you wish to delete
  • Parameters: None required
  • Output: Success message or empty message if file doesn't exist

Example Use Case: Cleaning up temporary files or removing files that are no longer needed.

Merge Workspace Item

Combines multiple workspace items into a single item.

Configuration:

  • Parameters:
    • Separator: The separator to use between merged items
  • Output: A new workspace item containing the merged content

Example Use Case: Combining multiple generated text fragments into a single document or concatenating files.

Move Directory

Moves a directory to a new location in the workspace.

Configuration:

  • Request: Source directory path
  • Parameters:
    • Destination: Target directory path
  • Output: Directory at new location

Example Use Case: Reorganizing workspace structure or relocating processed directories.

Move File

Moves a file to a new location in the workspace.

Configuration:

  • Request: Source file path
  • Parameters:
    • Destination: Target file path
  • Output: File at new location

Example Use Case: Reorganizing files after processing or moving files to designated output directories.

Split Workspace Item

Divides a workspace item into multiple items based on a separator.

Configuration:

  • Request: A workspace input reference to the document you wish to split
  • Parameters:
    • Separator: The separator used to split the item
  • Provided Parameters:
    • Current: The current split iteration (number)
  • Output: Multiple files, using the operation parameter "Current" to output multiple workspace files

Example Use Case: Breaking large documents into smaller sections for parallel processing.

Write File

Writes content to a file in the workspace. Can create new files or overwrite existing ones.

Configuration:

  • Request: Content to write to the file
  • Parameters:
    • File Path: Path where the file should be written
    • Encoding: Character encoding (default: UTF-8)
  • Output: Written file in workspace

Example Use Case: Saving generated content, configuration files, or processed data to the workspace.

Image

Create PDF

Converts content (typically Markdown) to a PDF document.

Configuration:

  • Request: Path to the markdown content you want to convert to PDF
  • Parameters: None required
  • Output: The generated PDF document

Example Use Case: Creating professional documentation or reports from markdown content generated by AI operations.

Execution

E2B: Execute Commands

Executes POSIX commands in an isolated E2B sandbox Linux VM.

Configuration:

  • Request: POSIX commands to be executed
  • Parameters:
    • Commands (required): The POSIX commands to execute
    • Timeout (optional): Maximum execution time in milliseconds
    • WorkspacePath (optional): Path to a workspace directory within the sandbox
  • Output: Standard output of the executed commands and workspace changes

Example Use Case: Running shell scripts or command-line tools in a controlled environment for tasks like data processing or system operations.

E2B: Execute Node.js

Executes Node.js code in an E2B sandbox.

Configuration:

  • Request: Node.js code to execute
  • Parameters:
    • Code (required): The Node.js code to execute
    • Timeout (optional): Execution timeout in milliseconds
    • NPMModules (optional): Comma-separated list of npm modules to install
    • WorkspacePath (optional): Working directory for execution
  • Output: Standard output from the Node.js process and workspace changes

Example Use Case: Executing Node.js scripts that require specific npm packages or complex JavaScript operations.

E2B: Execute Python

Executes Python code in an E2B sandbox.

Configuration:

  • Request: Python code to execute
  • Parameters:
    • Code (required): The Python code to execute
    • Timeout (optional): Maximum execution time in milliseconds
    • PIPPackages (optional): Comma-separated list of pip packages to install
    • WorkspacePath (optional): Working directory for execution
  • Output: Standard output of the Python code and workspace changes

Example Use Case: Running Python scripts that require specific pip packages or isolation from the main environment.

Execute JavaScript

Executes JavaScript or TypeScript code within the workspace context. Built-in Raikoo functions available.

Configuration:

  • Request: JavaScript or TypeScript code to execute
  • Parameters: None required
  • Output: The result of the executed code as a string

Available Functions: - raikoo.markdownToDocxAsync(): Convert Markdown to DOCX - raikoo.renderDocxTemplateAsync(): Render DOCX template with variables

Example Use Case: Performing custom data transformations, calculations, or integration with external services via APIs.

Execute Python

Executes Python code within the workspace context.

Configuration:

  • Request: Python code to execute
  • Parameters: None required
  • Output: The result of the last evaluated expression

Example Use Case: Leveraging Python libraries for data analysis, scientific computing, or machine learning tasks.

Misc

Error

Throws an error to halt workflow execution. Useful for validation failures or intentional workflow termination.

Configuration:

  • Request: Error message to display
  • Parameters: None required
  • Output: Workflow execution stops with error message

Example Use Case: Implementing validation checks that stop the workflow when critical conditions are not met.

Run Workflow

Executes a nested workflow for modular composition.

Configuration:

  • Parameters:
    • Workflow (required): Reference to the workflow to execute, optionally targeting a specific version
    • Additional parameters are passed as workflow parameters to the nested workflow
  • Output: The nested workflow's results are added to the workspace

Example Use Case: Creating reusable workflow modules that can be called from multiple parent workflows. For example, a document processing workflow that can be invoked by different business processes.

Select HTML

Extracts specific HTML elements using CSS selectors.

Configuration:

  • Request: The CSS selector to use (optional if "Selector" parameter is set)
  • Parameters:
    • HTML: The HTML document to select from
    • Selector: The CSS selector to use (takes precedence over request)
    • Format (optional): Set to "true" to format the extracted HTML
  • Output: An array of HTML elements matching the specified selector

Example Use Case: Targeted extraction of specific content from HTML documents, such as article bodies, tables, or navigation elements.

SendGrid Email

Sends emails using the SendGrid service. Configure recipients, subject, body, and attachments.

Configuration:

  • Request: Email body content
  • Parameters:
    • To: Recipient email address(es)
    • From: Sender email address
    • Subject: Email subject line
    • Attachments: Files to attach from workspace
    • CC/BCC: Additional recipients
  • Output: Delivery confirmation

Example Use Case: Sending automated reports, notifications, or generated documents via email to stakeholders.

Sleep

Pauses workflow execution for a specified duration. Useful for rate limiting or waiting between operations.

Configuration:

  • Request: Duration in milliseconds
  • Parameters:
    • Duration: Time to pause in milliseconds
  • Output: Resumes execution after delay

Example Use Case: Adding delays between API calls to respect rate limits, or waiting for external systems to complete processing.

SSH to Remote Linux

Executes scripts on remote Linux servers via SSH.

Configuration:

  • Parameters:
    • ResourceConnection (required): The Resource and SSH connection to use for remote execution
    • ScriptPath (optional): Path to a script file in the workspace to execute
    • Script (optional): Script content to execute directly
    • RemoteRootPath (optional): Root path on the remote server (default: /home/username/)
    • StdOutFile (optional): If provided, writes stdout to this filename
    • StdErrFile (optional): If provided, writes stderr to this filename
    • DoMountWorkspaces (optional): Automount persistent workspaces (requires root)
  • Output: Script execution results (stdout/stderr)

Example Use Case: Deploying applications, running maintenance scripts, collecting system information, or executing remote commands as part of DevOps workflows.

Write Workflow Parameter

Writes or updates a workflow parameter value during execution. Allows operations to pass data to subsequent operations via parameters.

Configuration:

  • Request: Value to write to the parameter
  • Parameters:
    • Parameter Name: Name of the parameter to write
  • Output: Updated parameter value available to subsequent operations

Example Use Case: Storing intermediate results or state that needs to be accessed by multiple operations throughout the workflow.

Git

Git: Add

Stages files for commit.

Configuration:

  • Request: Files to add (optional if "All" parameter is true)
  • Parameters:
    • GitPath (required): Path to the Git repository
    • All (optional): If true, all changed files are added
  • Output: Updated workspace reflecting the staging changes

Example Use Case: Staging changes as part of a Git workflow automation.

Git: Create Branch

Creates a new branch in the Git repository.

Configuration:

  • Request: Name of the new branch
  • Parameters:
    • GitPath (required): Path to the Git repository
  • Output: Repository with new branch created

Example Use Case: Creating feature branches as part of automated development workflows.

Git: Checkout Branch

Checks out a branch or ref.

Configuration:

  • Request: The branch or ref to checkout
  • Parameters:
    • GitPath (required): Path to the Git repository
  • Output: Updated workspace files reflecting the checkout operation

Example Use Case: Switching branches in automated testing or deployment workflows.

Git: Clone Repository

Clones a repo from URL.

Configuration:

  • Request: The URL of the Git repository to clone
  • Parameters:
    • GitPath (required): Path within the workspace for the cloned repository
  • Output: The cloned Git repository in the workspace

Example Use Case: Pulling external code or content repositories into a workflow for processing.

Git: Commit

Commits staged changes.

Configuration:

  • Request: The commit message
  • Parameters:
    • GitPath (required): Path to the Git repository
    • Name (required): Author name for the commit
    • Email (required): Author email for the commit
    • Message (optional): Commit message (if not provided in request)
  • Output: Updated workspace reflecting the committed changes

Example Use Case: Automating version control as part of content generation or code modification workflows.

Git: Init

Initializes new repo.

Configuration:

  • Request: Path to initialize the repository (optional if GitPath is set)
  • Parameters:
    • GitPath (optional): Path to initialize the repository
    • DefaultBranch (optional): Name of the default branch (defaults to "main")
    • Bare (optional): Whether to create a bare repository
  • Output: Newly initialized Git repository

Example Use Case: Setting up version control for newly generated projects or content.

Git: Pull

Pulls from remote.

Configuration:

  • Request: Path to the Git repository
  • Parameters:
    • GitPath (required): Path to the Git repository
    • Remote (optional): Remote repository to pull from (defaults to "origin")
    • Branch (optional): Local branch to pull into (defaults to current branch)
    • RemoteBranch (optional): Remote branch to pull from
  • Output: Updated workspace reflecting the pulled changes

Example Use Case: Updating local content from a remote repository before processing.

Git: Pull Request (github)

Creates a GitHub pull request.

Configuration:

  • Request: Pull request body/description
  • Parameters:
    • Repository: GitHub repository (owner/repo format)
    • Base Branch: Target branch for the PR
    • Head Branch: Source branch with changes
    • Title: PR title
    • Draft: Whether to create as draft PR
  • Output: Created pull request URL and details

Example Use Case: Automating PR creation for generated code changes or documentation updates.

Git: Push

Pushes to remote.

Configuration:

  • Request: A workspace containing the Git repository
  • Parameters:
    • GitPath (required): Path to the Git repository
    • Remote (required): Name of the remote repository
    • Branch (required): Name of the local branch to push
    • RemoteBranch (optional): Name of the remote branch to push to
  • Output: Updated workspace reflecting any changes from the push operation

Example Use Case: Publishing generated content or code modifications to a remote repository.

Git: Reset Hard

Hard resets to HEAD.

Configuration:

  • Request: Path to the Git repository
  • Parameters:
    • GitPath (required): Path to the Git repository
  • Output: Updated workspace reflecting the repository state after reset

Example Use Case: Reverting to a clean state when operations need to start fresh.

RAG

Vector Store Clear Collection

Clears all documents from a vector store collection without deleting the collection itself.

Configuration:

  • Parameters:
    • Connection: Vector store connection to use
    • Collection Name: Name of the collection to clear
  • Output: Confirmation of cleared collection

Example Use Case: Resetting a collection before reindexing with updated documents.

Vector Store Create Collection

Creates a new collection in the configured vector store.

Configuration:

  • Parameters:
    • Connection: Vector store connection to use
    • Collection Name: Name for the new collection
    • Embedding Model: Model to use for generating embeddings
    • Dimension: Vector dimension (depends on embedding model)
  • Output: Created collection ready for document loading

Example Use Case: Setting up a new knowledge base for RAG-powered AI operations.

Vector Store Delete Collection

Permanently deletes a collection and all its documents from the vector store.

Configuration:

  • Parameters:
    • Connection: Vector store connection to use
    • Collection Name: Name of the collection to delete
  • Output: Confirmation of deletion

Example Use Case: Removing obsolete knowledge bases or cleaning up test collections.

Vector Store Load Documents

Loads and indexes documents into a vector store collection for RAG retrieval.

Configuration:

  • Parameters:
    • Connection: Vector store connection to use
    • Collection Name: Target collection
    • Documents: Files or text to index
    • Chunk Size: Size of text chunks for embedding
    • Chunk Overlap: Overlap between chunks
    • Metadata: Additional metadata to store with documents
  • Output: Indexed documents ready for retrieval

Example Use Case: Building a searchable knowledge base from documentation, articles, or internal documents.

Vector Store Query

Queries a vector store collection with a text query, returning relevant document chunks.

Configuration:

  • Parameters:
    • Connection: Vector store connection to use
    • Collection Name: Collection to query
    • Query: Text query to search for
    • Top K: Number of results to return
    • Score Threshold: Minimum similarity score
  • Output: Relevant document chunks with similarity scores

Example Use Case: Finding relevant information from a knowledge base based on semantic similarity.

Vector Store Retrieval

Performs RAG retrieval, querying the vector store and formatting results for use as AI context.

Configuration:

  • Parameters:
    • Connection: Vector store connection to use
    • Collection Name: Collection to retrieve from
    • Query: Search query
    • Top K: Number of chunks to retrieve
    • Format Template: How to format results for AI context
  • Output: Formatted context ready to be included in AI prompts

Example Use Case: Augmenting AI operations with relevant context from a knowledge base to improve response accuracy.

Best Practices

Operation Selection: - Choose the right operation category for your task - Use Agent operations for complex multi-step reasoning tasks - Use system operations for deterministic, programmatic tasks - Leverage RAG operations when AI needs access to large knowledge bases

Error Handling: - Validate inputs before processing - Use the Error operation for explicit validation failures - Handle operation failures gracefully with conditional logic

Performance: - Minimize large file operations - Use appropriate sandboxing (E2B) for untrusted code - Cache expensive operation results in the workspace - Consider timeouts for long-running operations

Modularity: - Create reusable workflows with Run Workflow operation - Use Split and Merge operations for parallel processing - Organize complex logic into separate workflow modules

Security: - Use E2B sandbox operations for executing untrusted code - Validate external resources when fetching from URLs - Store sensitive credentials in connection configurations, not in code - Review AI-generated code before execution