Tensorify LogoTensorify
Tensorify LogoTensorify
Log inGet Started

Guides

Step-by-step walkthroughs for building real workflows. Each guide results in a deployed, working automation by the end.

Quick-Start Templates

Ready-to-use workflows you can create with one click — Echo Server, Hello World API, API Proxy, JSON Reshaper, Conditional Alert, Data Enrichment, and Custom Python Processor.

Quick-Start Templates — Seven beginner-to-intermediate templates covering webhooks, data transformation, conditional routing, API integration, and custom Python processing.

  • Uses: Webhook Trigger, Transform, HTTP Request, If, Code, Return, Stop
  • Time: ~5 minutes per template

Getting Started

If this is your first workflow, start here:

Send Email on Trigger — The simplest useful workflow. Run it manually and it sends a personalized email. Good for verifying your full setup end-to-end.

  • Uses: Manual Trigger, Code, Send Email
  • Time: ~10 minutes

Automation Workflows

Webhook Automation — Receive a GitHub push event, extract commit data, and send a team notification email with HMAC verification.

  • Uses: Webhook Trigger, Transform, Code, Send Email
  • Time: ~15 minutes

API Workflows

Build a REST API Endpoint — Expose an authenticated HTTP endpoint that accepts a POST request, calls an external service, and returns a JSON response synchronously.

  • Uses: API Endpoint Trigger, HTTP Request, Transform, Return
  • Time: ~10 minutes

Advanced

Chain Workflows with Subworkflow — Compose a pipeline from three separate workflows — a parent that calls two child workflows in sequence.

  • Uses: Webhook Trigger, Subworkflow, Manual Trigger, If, Return
  • Time: ~20 minutes

AI Agents

Build an AI Agent — Build, test, and deploy an autonomous AI agent with tool calling, memory, and OpenAI SDK compatibility.

  • Uses: AI Agent, HTTP Request, Window Memory, API Endpoint, Return
  • Time: ~15 minutes

Agent Memory — Choose and configure memory for AI agents — Window Memory vs Qdrant Memory, session key patterns, and production tips.

  • Uses: AI Agent, Window Memory, Qdrant Memory
  • Time: ~10 minutes

Self-Host an AI Agent — Deploy an AI agent on your own machine using the CLI runner. Use local LLMs via Ollama, access local files, and keep data private.

  • Uses: API Trigger, AI Agent, Window Memory, Return, CLI Runner
  • Time: ~15 minutes

Deploy as an OpenAI Endpoint — Turn any workflow into an OpenAI-compatible API endpoint. Connect from any OpenAI SDK with streaming and session memory.

  • Uses: API Trigger, AI Agent, Return
  • Time: ~10 minutes

Process Local Files with an AI Agent — Build an AI agent that reads and analyzes local files. Create a codebase assistant running on your machine.

  • Uses: API Trigger, AI Agent, Code (tools), Return, CLI Runner
  • Time: ~15 minutes

Build a RAG System — Build a retrieval-augmented generation system with Qdrant vector search and an AI agent.

  • Uses: API Trigger, AI Agent, Qdrant Memory, Code, Return
  • Time: ~20 minutes

Streaming AI Responses — Stream real-time token-by-token responses from AI agents using Server-Sent Events. Works with OpenAI SDKs, Vercel AI SDK, and browser JavaScript.

  • Uses: AI Agent, API Trigger or Webhook Trigger, SSE streaming
  • Time: ~10 minutes

Trigger Selection

Choosing a Trigger — API Trigger, Webhook Trigger, or Manual Trigger — which to use when. Opinionated guidance with a decision table.

  • Covers: All trigger types, AI agent-specific recommendations
  • Time: ~5 minutes

REST APIs

Build a CRUD API — Build a full REST API with GET, POST, PUT, DELETE operations using the visual editor. No framework needed.

  • Uses: API Trigger, Switch, DB Find, DB Insert, DB Update, DB Delete, Return
  • Time: ~15 minutes

Messaging & Chatbots

Build a Messaging Chatbot — Build an AI chatbot on Telegram, Slack, or Discord with conversation memory and automatic replies.

  • Uses: Telegram/Slack/Discord Trigger, AI Agent, Memory, Send Message
  • Time: ~10 minutes

MCP (Model Context Protocol)

Expose a Workflow as an MCP Tool — Turn any workflow into a callable tool that AI agents (Claude, Cursor, custom) can discover and invoke programmatically.

  • Uses: MCP Trigger, any action/logic nodes, Return
  • Time: ~10 minutes

Tensorify MCP Server for AI IDEs — Connect Cursor or Claude Desktop to Tensorify's MCP Platform Server. Let your AI agent create, edit, deploy, and run workflows without opening the browser.

  • Uses: CLI tensorify mcp config, API Key, MCP Platform (20 tools)
  • Time: ~5 minutes

Local Machine & DevOps

Local Machine Automation — Use File Ops, Shell, and System Info plugins to automate tasks on your local infrastructure via the CLI runner.

  • Uses: File Ops, Shell, System Info, AI Agent, CLI Runner
  • Time: ~15 minutes

Build a Coding Agent — Create an AI coding assistant that reads files, writes code, runs tests, and streams responses via OpenAI-compatible API.

  • Uses: API Trigger, AI Agent, File Ops, Shell, System Info, Web Search, Memory, Return
  • Time: ~20 minutes

All guides assume you have a Tensorify workspace set up. If you haven't done that yet, follow the Quick Start first.

On this page