Custom neural network pipelines are now available in PIPER. You no longer need to hire programmers — you can easily create an Custom AI pipelines yourself.
Previously, solving non-standard tasks required developers, weeks of waiting, and significant costs. Now you can build custom pipelines visually in Piper’s editor and launch them instantly.
What Is a Custom Pipeline?
A custom AI pipeline is a set of neural network blocks connected visually in Piper by dragging and dropping, no code required.
Think of it like LEGO: each block easily connects to another, except the blocks are neural networks.
For example, your website or app might require image processing: background removal or quality enhancement. Previously, you’d need developers, technical specifications, and significant expenses.
Now, such features are quickly assembled in Piper’s visual editor: add a block to remove backgrounds, another to enhance image quality, and finally, generate descriptions for your images.
For typical tasks, use pre-built pipelines that don’t require assembly.
Real-World Examples of Custom Pipelines
Here are concrete situations where custom pipelines benefit Piper users:
Online clothing store. Product images automatically process by removing backgrounds, adjusting colors, and increasing resolution. Variations with different styles and colors are generated instantly.
Unique avatar generator. Users upload a photo, and the neural network applies a style, adds characteristic features, and generates a finished image immediately after processing.
Telegram horoscope bot. Users input their birthdate, and the pipeline generates horoscope text and selects corresponding images, automatically updating every morning.
Customer support chatbot. Handles typical queries like order status or delivery times, automatically routing complex requests to a manager via email or messenger.
Automated article illustrations: The pipeline analyzes article text and generates images based on key phrases, eliminating the need to find or purchase stock photos.
How to Build Your Pipeline in Piper
Go to PIPER
Click «Create» to start a new pipeline
To add blocks, click «+»
Select a required block, which automatically appears on the Canvas
Add another block via «+» and connect the pipelines
Done! Click Play to launch your pipeline. The cost per use is displayed nearby.
Rename or change the pipeline version in the «Design» टैब.
To delete a pipeline, go back to Projects, click the burger menu next to the pipeline name, and select Delete.
Here’s a practical example: creating a Telegram horoscope bot.
Start with the user’s birthdate input, adding the Input Text node.
- Next, add an LLM node like ChatGPT or Claude. Insert a prompt, for example:
«Determine the zodiac sign based on birthdate and write a positive horoscope. Additionally, create a prompt to generate an image with the horoscope and zodiac sign for FLUX»
- Then, use an Extract JSON node to separately handle the response and FLUX request.
- Add free image generation using the FLUX model.
All that’s left is creating a bot via @BotFather and linking the pipeline through API. How to use Piper’s API
You can generate bot code directly in Piper. It might look like this:
import requests
from telegram import Update
from telegram.ext import ApplicationBuilder, CommandHandler, MessageHandler, ContextTypes, filters
PIPER_API = "https://api.piper.my/run/{pipeline_id}" # YOUR URL
async def start(update: Update, context: ContextTypes.DEFAULT_TYPE): await update.message.reply_text("Hi! Send me your birth date to see the horoscope 😊")
async def handle_message(update: Update, context: ContextTypes.DEFAULT_TYPE):
date = update.message.text.strip()
await update.message.reply_text("Generating your horoscope ✨")
payload = {"input_text": date}
response = requests.post(PIPER_API, json=payload)
if response.status_code == 200:
result = response.json()
image_url = result["outputs"]["image"]
await update.message.reply_photo(photo=image_url, caption=text)
else:
await update.message.reply_text("Oops.. Something went wrong 😢")
app = ApplicationBuilder().token("YOUR_TELEGRAM_BOT_TOKEN").build()
app.add_handler(CommandHandler("start", start))
app.add_handler(MessageHandler(filters.TEXT & ~filters.COMMAND, handle_message))
app.run_polling()
What Else Can Piper Do?
Piper includes dozens of nodes: image generation, video enhancement, text processing, photo editing, and more.
Each block works individually or within a pipeline, allowing you to insert, replace, or rearrange them as needed.
Find all available modules here.
Why Use Custom Pipelines?
Custom pipelines eliminate the need to hire developers, significantly reducing the time from idea to launch—features can now be tested and implemented within hours instead of weeks, while also automating routine operations and freeing your team from manual data processing.
Why Choose Piper?
Here’s why users prefer creating custom pipelines with Piper:
Easy and intuitive visual editor — build complex AI products without coding by simply dragging blocks.
Instant experimentation and testing — no need to wait for programmers or spend money on every minor change.
Simple scalability — your pipelines easily grow with your business.
Reliable infrastructure — Piper handles millions of requests daily, maintaining stable and uninterrupted performance.
Access to popular AI technologies in one place.
Try It Yourself
Don’t take our word for it — visit Piper now and build your first pipeline in minutes.
We give all new users $1 credit, enough for 1000 image upscales or 100 FLUX generations.