Components
In SmythOS, components are the building blocks of your agent workflows. Each one represents a single task, skill, or utility, like calling an API, running logic, filtering JSON, generating text, or storing knowledge.
You build workflows by dragging components onto the canvas, connecting them visually, and configuring how data flows between them.
Designing Logic with Components
Each pattern helps you control how your agent thinks and acts. You can combine components like Classifier, ForEach, or Async/Await to create dynamic, multi-step behaviors without writing code.
Sequential Flow
Run steps one after another.
Use when: Each step depends on the output of the previous
Example: Generate blog → Summarize → Post to site
Useful components:
Branching Flow
Choose a path based on logic or classification.
Use when: Your workflow must adapt to input data
Example: If sentiment is negative, alert support; if positive, publish
Useful components:
Parallel Flow
Run multiple tasks at the same time.
Use when: Tasks don’t depend on each other
Example: Analyze text and extract entities in parallel
Useful components:
Choosing the Right Component
What You Want to Do | Use This Component | Why It Helps |
---|---|---|
Loop through a list | ForEach | Handles each item one by one |
Run tasks at the same time | Async/Await | Runs branches in parallel |
Add a timed pause | Sleep | Controls timing or rate limits |
Store useful facts or context | RAG Remember | Saves info for later use in the workflow |
Route based on input type | Classifier | Classifies input and picks the right path |
Transform or filter JSON | JSON Filter | Modifies or extracts specific JSON data |
Base Components
These are the essential building blocks that help your agents interact, route input, and generate output.
Agent Skill
Define custom functions or actions your agent can call on demand.
Classifier
Categorize inputs using AI classification to support decision logic.
GenAI LLM
Prompt large language models to generate intelligent responses.
Image Generator
Create AI-generated visuals from prompts using models like DALL·E.
Note
Add non-functional comments and instructions inside your workflows.
Advanced Components
These components support more complex logic, loops, filters, and low-level execution.
API Call
Connect to third-party services or internal APIs to fetch or send data.
ForEach
Loop over arrays or lists to perform repeated tasks.
JSON Filter
Transform, extract, or restructure JSON responses with precision.
Code
Run JavaScript logic inline using Node.js and custom logic.
Async
Run multiple branches in parallel for concurrent processing.
Await
Pause a workflow until all async branches have completed.
LLM Assistant
Use large language models to generate text, analyze inputs, or assist with decision-making.
Sleep
Pause execution between tasks for better timing control.
Crypto Components
Secure operations for signing, hashing, and transforming data formats.
Encode/Decode
Convert between formats like Base64 or UTF-8.
Sign
Generate digital signatures and verify payload integrity.
Hash
Hash strings securely using algorithms like SHA256.
Timestamp
Add secure or formatted timestamps to data or logs.
RAG (Retrieval-Augmented Generation) Components
Memory-aware components for knowledge management and dynamic search.
RAG Remember
Store custom facts or content for retrieval later.
RAG Search
Search stored knowledge using semantic queries.
RAG Forget
Delete or update parts of your agent's knowledge.
Tool Components
Powerful utilities to connect with browsers, scripts, and live data.
Web Scrape
Extract and parse data from public websites and URLs.
Web Search
Query search engines and receive summarized results.
Computer Use
Trigger commands and automate desktop-level interactions.
MCP Client
Send data or trigger workflows to your external MCP setup.
Node.js
Run native JavaScript inside the workflow.
Why Components Matter
Best Practices
- Use clear component names and Notes to document your logic
- Start simple, then layer in logic like loops or conditionals
- Test individual components before combining them into larger flows
- Use Agent Skills to break big flows into manageable pieces
- For memory, use RAG components to store reusable facts
What to Try Next
- Explore the Workflow Guide to see how components come together
- Learn how to Deploy Your Agent
- Use a ready-made Agent Template to build faster