Quick-Start Templates

Every template below is a working workflow you can deploy in under five minutes. Click the "Create this workflow" link to open Tensorify with the nodes already wired up — then hit Test to see live output.

Prerequisites:

• A Tensorify account with a workspace
• CLI installed if you plan to deploy locally: curl -fsSL https://cli.tensorify.io/install | sh (Windows: irm https://cli.tensorify.io/install.ps1 | iex)

Difficulty: Beginner · Nodes: Webhook Trigger → Return

The simplest possible workflow. It receives any webhook request and returns the full envelope — body, headers, query string, HTTP method, and path. Use it to inspect what external services are actually sending you.

When to use this:

• Debugging a third-party webhook (Stripe, GitHub, Shopify) to see the exact payload
• Building a health-check endpoint for monitoring tools
• Learning how Tensorify nodes connect

Difficulty: Beginner · Nodes: Webhook Trigger → Transform → Return

Takes a name from the request body and returns a personalised greeting. Demonstrates the Transform node's {{ }} binding syntax, which lets you build JSON responses from incoming data without writing code.

When to use this:

• First-time introduction to Tensorify's data-binding syntax
• Quick prototype of a micro-API that formats a response from user input

Example request:

{ "name": "Faisal" }

Example response:

{ "greeting": "Hello, Faisal!", "receivedAt": "2026-05-15T17:28:37Z" }

Difficulty: Beginner · Nodes: Webhook Trigger → HTTP Request → Return

Forwards an incoming webhook request to an external REST API and returns the result. The default calls JSONPlaceholder, but you can swap the URL to proxy any public API — weather, exchange rates, or your own internal services.

When to use this:

• Adding CORS headers or auth to an API your frontend can't call directly
• Aggregating multiple APIs behind a single Tensorify endpoint
• Quickly testing external API responses without writing a server

Difficulty: Beginner · Nodes: Webhook Trigger → Transform → Return

Takes flat, inconsistent JSON and remaps it into a clean, standardised structure. The example converts raw contact data (separate first_name, last_name, email_address fields) into a unified contact object.

When to use this:

• Normalising webhook payloads from different CRMs or forms before storing them
• Mapping between two API formats (e.g., Salesforce → HubSpot)
• Cleaning up data before passing it to a downstream workflow

Example input:

{
  "first_name": "Faisal",
  "last_name": "Sifat",
  "email_address": "[email protected]",
  "phone": "+1234567890"
}

Example output:

{
  "contact": {
    "fullName": "Faisal Sifat",
    "email": "[email protected]",
    "phone": "+1234567890"
  },
  "source": "webhook",
  "processedAt": "2026-05-15T17:30:17Z"
}

Difficulty: Intermediate · Nodes: Webhook Trigger → If → Transform / Stop → Return

Routes incoming events based on a condition. High-priority events build a critical alert response; low-priority events are silently dropped with the Stop node.

When to use this:

• Filtering noisy monitoring alerts — only forward critical ones to Slack / PagerDuty
• Building a simple triage layer for event-driven systems
• Demonstrating branching logic to your team

High-priority response:

{
  "alert": true,
  "severity": "critical",
  "message": "Server CPU at 95%",
  "action": "Notify on-call engineer"
}

Difficulty: Intermediate · Nodes: Webhook Trigger → HTTP Request → Transform → Return

Receives a request, fetches additional data from an external API, and merges the results into a clean response. The example calls JSONPlaceholder to look up a user profile and returns only the essential fields.

When to use this:

• Enriching CRM leads with company data from Clearbit, Apollo, or your own database
• Building an internal lookup service that combines multiple data sources
• Adding computed fields (like enriched: true or a timestamp) before returning data

Example output:

{
  "user": {
    "name": "Leanne Graham",
    "email": "[email protected]",
    "company": "Romaguera-Crona"
  },
  "enriched": true,
  "source": "jsonplaceholder-api"
}

Difficulty: Intermediate · Nodes: Webhook Trigger → Code → Return

Runs arbitrary Python on the incoming data. The example filters a list of numbers above a threshold and computes statistics. Use this when Transform bindings aren't enough and you need loops, math, or custom logic.

When to use this:

• Running business logic that's too complex for Transform expressions
• Batch-processing arrays, computing aggregations, or validating schemas
• Prototyping a data pipeline step before moving it to production code

Example request:

{ "items": [10, 25, 3, 47, 8, 15], "threshold": 10 }

Example response:

{
  "filtered": [25, 47, 15],
  "count": 3,
  "total": 6,
  "average": 29,
  "max": 47
}

Difficulty: Beginner · Category: AI Agents · Nodes: API Trigger (openai-chat) → AI Agent → Return

Deploy a chatbot as an OpenAI-compatible endpoint. Call it from any OpenAI SDK — Python, Node.js, or curl. This is the fastest way to ship an AI agent.

When to use this:

• Deploying a chatbot that any OpenAI SDK can connect to
• Building a custom AI assistant for your application
• Starting point for more complex AI workflows

Required secrets: OPENAI_API_KEY

Difficulty: Intermediate · Category: AI Agents · Nodes: API Trigger (openai-chat) → Qdrant Memory → AI Agent → Return

A knowledge base that retrieves relevant documents from Qdrant and uses an AI agent to answer questions with context. Deploy as an OpenAI-compatible endpoint for document Q&A.

When to use this:

• Building a documentation bot that answers questions from your own content
• Creating a product FAQ assistant with semantic search
• Internal knowledge base for teams

Required secrets: OPENAI_API_KEY

See the full Build a RAG System guide.

Difficulty: Intermediate · Category: AI Agents · Nodes: API Trigger (openai-chat) → Window Memory → AI Agent (Ollama) → Return

Run an AI agent on your own machine using Ollama. No cloud API keys, no data leaving your infrastructure. Requires Ollama running locally.

When to use this:

• Running AI agents with zero API costs
• Processing sensitive data that can't leave your network
• Testing different open-source models locally

See the Self-Host an AI Agent guide.

Difficulty: Beginner · Category: APIs · Nodes: API Trigger (REST) → Switch → DB Find / Insert / Update / Delete → Return

A full REST API backend with GET, POST, PUT, and DELETE operations on a database table. No framework, no boilerplate — just visual nodes.

When to use this:

• Building a backend for a web or mobile app
• Prototyping an API before writing server code
• Replacing a simple Express/Django CRUD server

Required secrets: DATABASE_URL

See the Build a CRUD API guide.

Difficulty: Beginner · Category: Webhooks · Nodes: Webhook Trigger (GitHub) → Transform → HTTP Request (Slack) → Return

Receive GitHub push events with automatic signature verification, extract commit details, and post a formatted notification to Slack.

When to use this:

• Getting Slack notifications for GitHub pushes, PRs, or issues
• Building a custom CI/CD notification pipeline
• Connecting any webhook source to Slack

Required secrets: WEBHOOK_SECRET

Difficulty: Intermediate · Category: AI Agents · Nodes: Webhook Trigger → AI Agent → DB Insert → Return

Use an AI agent to extract structured information from raw data and store it in a database. The agent interprets unstructured input and returns clean, typed JSON.

When to use this:

• Extracting company data from emails or form submissions
• Enriching CRM leads with AI-powered analysis
• Processing documents into structured database records

Required secrets: OPENAI_API_KEY, DATABASE_URL

Your own AI coding assistant — reads files, writes code, runs tests, and iterates until passing. Streams responses via OpenAI-compatible API. Requires CLI runner.

Required secrets: OPENAI_API_KEY, SEARCH_API_KEY

AI chatbot on Telegram — receives messages, thinks with conversation memory, and replies automatically. Deploy and connect to @BotFather in minutes.

Required secrets: TELEGRAM_BOT_TOKEN, OPENAI_API_KEY

Mention the bot in Slack to run deployments, check server health, or execute commands. AI-powered DevOps assistant running on your infrastructure.

Required secrets: SLACK_BOT_TOKEN, SLACK_SIGNING_SECRET, OPENAI_API_KEY

Also available in the AI Agents category on the dashboard:

• AI Chatbot (Manual) — the simplest agent workflow: Manual Trigger → AI Agent → Return
• AI Agent with Tools — an agent that can call HTTP APIs and run custom code
• Customer Support Agent — a support agent with conversation memory and API access

• Deploy a workflow: Once you've tested your template, deploy it as a live webhook endpoint.
• Self-host: Deploy an AI agent on your own machine with the CLI runner.
• Build a RAG system: Follow the Build a RAG System guide.
• Chain workflows: Use the Subworkflow plugin to call one workflow from another.
• Add secrets: Store API keys safely with Environment Variables.
• Explore all plugins: See the full Plugin Reference.