2,484Book a call

Open-source automation for construction data

Workflow automation for construction with n8n, Dify, Flowise and LangFlow

Visual, open-source platforms that build complex workflows without code, much like Dynamo or Grasshopper. With our ready pipelines you can automate CAD/BIM data processing, validation, QTO and cost estimation today.

  • Windows, Mac and Linux
  • Offline or online
  • First workflow in under 30 minutes
See ready pipelines Build it with our team
cad-bim-pipeline.jsonrunning locally
  1. Model fileRVT · IFC · DWG
  2. Convertcad2data
  3. ValidateExcel rules
  4. ClassifyAI agent
  5. ReportXLSX · HTML
Build it in n8n Dify Flowise LangFlow

Why automate

Tools exist. Data chaos keeps people clicking.

A major challenge in using new technologies is that data, while abundant, remains fragmented, unstructured, and often incompatible between different systems and programs. Tools exist, but data chaos keeps the industry trapped in manual processes. The solution is automation, accessible and understandable for everyone, not just programmers.

  • Open sourceNo vendor lock-in.
  • Hundreds of integrationsReady nodes for files, mail, databases, APIs and AI.
  • Custom nodesWrite your own in JavaScript or Python.
  • Your infrastructureDeploy on your own servers or a laptop.
  • GDPR-friendlyData stays in local storage.
  • Fair-code licencen8n can be self-hosted under its fair-code licence.

Open-source platforms

AI workflow and automation platforms, compared.

Choose the right tool for the job. All four are open source and can run on your own machine.

n8n

Workflow automation

Visual workflow automation with 400+ integrations. Ideal for file operations, APIs, notifications and data processing.

  • 400+ nodes
  • Self-hosted
  • Fair-code
Official site

Dify

Agentic AI platform

Build AI agents and RAG pipelines visually. Supports 100+ LLM models, document processing and autonomous reasoning.

  • RAG pipeline
  • AI agents
  • Open source
Official site

Flowise

LLM workflow builder

Drag-and-drop LLM workflows with 100+ integrations. Multi-agent orchestration and Graph RAG.

  • Multi-agent
  • Graph RAG
  • Open source
Official site

LangFlow

Visual AI builder

Python-based visual builder by DataStax. Native LangChain integration, MCP support, rapid prototyping.

  • LangChain
  • MCP
  • Open source
Official site

Prefer to describe a task in words instead of drawing it? AI agents for construction

n8n essentials

What makes n8n different: features everyone should know.

1. Visual workflow editor

A canvas where you connect nodes for services and actions. Every node shows its inputs and outputs, so the data flow is visible.

2. Error handling and debugging

n8n is strong at error management:

  • Error workflows: separate workflows that start when a main one fails
  • Retry logic: retry failed runs at set intervals
  • Debug mode: step through a workflow node by node
  • Execution history: past runs with full data snapshots

3. Advanced features

For power users:

  • Webhooks: custom endpoints that trigger workflows from other systems
  • Schedules: cron or fixed intervals
  • Sub-workflows: modular, reusable building blocks
  • Expressions: JavaScript for complex transformations
  • Split and merge: process in parallel, then combine

4. Security and compliance

Critical for sensitive project data:

  • Encryption: credentials encrypted at rest
  • Role-based access: control who views and edits
  • Audit logs: track changes and executions
  • On-premise: all data inside your infrastructure

Quick start

Install n8n in three steps.

Then take a ready-made pipeline and launch it. For agents in code, start with DDC Skills for AI Agents on GitHub.

  1. 1

    Install Node.js

    From the official site. Takes about two minutes.

  2. 2

    Install n8n with one command

    In CMD or PowerShell:

    npx n8nornpm install n8n -g
  3. 3

    Open n8n and start building

    In your browser at localhost:5678. Import a workflow JSON and press Execute.

Four steps: install Node.js, start n8n, download the repository, import a workflow
n8n quick start: installation and your first pipeline

AI integration

Let a language model write the pipeline.

Claude, ChatGPT and other models plug straight into n8n: add the node and write the prompt for text generation, data analysis or reports. Modern models can also generate a working n8n pipeline in minutes from a description of the logic.

Example prompt

Create a JSON pipeline for n8n with the following logic: 1. Read files from a folder 2. Extract data from PDFs 3. Save to Excel with formatting 4. Send report via email with attachments

  • OpenAIChatGPT, DALL-E, Whisper
  • AnthropicClaude models for advanced reasoning
  • Google AIGemini API
  • Hugging FaceOpen-source models
  • Local LLMsSelf-hosted models via Ollama
An LLM writes the n8n workflow JSON, which is imported and run in n8n
Automate CAD/BIM workflows locally with n8n + ChatGPT & Claude

Templates are imported as JSON, adapted and run immediately. Data from CAD, BIM, CAFM, ERP, Excel and cloud services starts speaking the same language, and automation becomes a tool for everyone, from project managers to design engineers.

Practical examples

CAD/BIM process automation, ready to download.

Every pipeline is free on GitHub and as a template on n8n.io. To use one: import the JSON, set the converter path and the project file path, press Execute.

01

Data conversion: Revit, IFC, DWG to a database and geometry

Converts .rvt, .ifc or .dwg files into an element database (.xlsx, .csv) and exports 3D geometry (.dae). Runs locally, without internet or Autodesk licences.

n8n_Revit_IFC_DWG_Conversation_simple.json

n8n template JSON on GitHub
n8n workflow: Data conversion: Revit, IFC, DWG to a database and geometry
02

CAD/BIM data quality validation

Reads CAD/BIM data and checks for missing or wrong parameters by group, type or any other criterion. The report shows completion percentages and unique parameter values; rules can come from your EIR, BEP or IDS.

n8n_Validation_CAD_BIM_Revit_IFC_DWG.json

n8n template JSON on GitHub
n8n workflow: CAD/BIM data quality validation
03

Quantity takeoff and report generation

Calculates volumes and other quantities for selected element groups and writes an HTML report. Made for quick grouping, calculations and estimates.

n8n_CAD_BIM_Quantity_TakeOff_HTML_Report_Generator.json

n8n template JSON on GitHub
n8n workflow: Quantity takeoff and report generation
04

Construction cost estimation for CAD (BIM) with an LLM

Estimates the cost of building elements from CAD/BIM files. AI classifies the materials, searches prices and builds a cost report.

n8n_5_Construction_Price_Estimation_with_LLM_for_Revt_and_IFC.json

n8n template JSON on GitHub
n8n workflow: Construction cost estimation for CAD (BIM) with an LLM

All 15 published workflows

Our n8n.io creator page
  • Project management bot with tasks and photo reportsA Telegram bot for task management and photo reports from the construction site.n8n.ioGitHub
  • Cost estimate from a Revit or IFC model with an LLMReads the model, groups elements and lets GPT or Claude price them line by line.n8n.ioGitHub
  • Validate CAD/BIM files against an Excel requirements tableChecks every element against your rules and returns a colour-coded report.n8n.ioGitHub
  • Convert Revit, IFC and DWG to tablesThe simple start: model in, XLSX out, plus DAE geometry for Revit and IFC.n8n.ioGitHub
  • ETL: extract model data and parse the XLSXConvert and parse into clean structured data, ready for LLM tasks.n8n.ioGitHub
  • Quantity takeoff as an interactive HTML reportWall volumes by type and other quantities, as a report you can open in any browser.n8n.ioGitHub
  • Estimate from text, photo or PDF in TelegramSend a message, a site photo or a drawing and get an estimate built on the open cost database.n8n.ioGitHub
  • Converter with all export settingsChoose 309, 724 or all 1,209 categories, bounding boxes, schedules and PDF sheets.n8n.ioGitHub
  • Batch converter pipelineWhole folders at once, with validation, timings and an HTML summary.n8n.ioGitHub
  • Carbon footprint (A1–A3) from Revit and IFCEmbodied carbon per element and material, with a report.n8n.ioGitHub
  • 4D/5D cost estimation pipeline from a BIM modelQuantities from the model, matched to work items and resources, with time and cost.n8n.ioGitHub
  • Estimate bot for text messagesDescribe the work in plain language in Telegram and get a priced breakdown.n8n.ioGitHub
  • Estimate from a site photoA vision model recognises the work in a photo and prices it.n8n.ioGitHub
  • AI classification with LLM and RAGMap elements to Omniclass, Uniclass or your own codes using a vector search over the cost database.GitHub
  • CAD/BIM to code: an LLM writes the Python for youDescribe the grouping, chart or QC report you need; the workflow writes and runs the code.GitHub

Auto classification and mapping

Classify and map elements automatically.

Pipelines sort building elements from Revit, IFC or DWG into standard categories and map their attributes for further calculations: quantities, cost, carbon footprint.

Element classification with LLM and RAG
  • Rough estimates, fasterWhen a cost or time estimate may carry an acceptable error margin, the manual element checks usually done by BIM coordinators and BIM managers are no longer needed.
  • Classification rules applied automaticallyRules that estimators apply by hand to elements and groups run inside the pipeline, which frees experts for strategic work.
  • New skills, new opportunitiesFrom engineers to estimators, teams move from reviewing and mapping data by hand to decisions and value creation.

ETL in code

What an ETL pipeline looks like in Python.

When a company is not ready for a full ERP, automation starts outside it: modular code that turns data into an open structured form, then runs checks, transformations and messages on a schedule. Pandas is the main tool.

Input, pipeline, output: Charts from Excel
Manual process
~5 minutes
Pipeline runtime
~5 seconds

  1. 1Extract open the data

    import pandas as pd
import matplotlib.pyplot as plt

# Table produced by the converter
DDCoutput_file = file_path[:-4] + "_rvt.xlsx"
df = pd.read_excel(DDCoutput_file)

  2. 2Transform group and visualise

    # Sum 'Amount' by 'Category' and plot it
ax = df.groupby('Category')['Amount'].sum().plot(
    kind='bar', figsize=(10, 5), color='skyblue',
    title='Expenses by Category', ylabel='Amount',
    rot=45, grid=True).get_figure()

  3. 3Load export

    # Save the chart as PNG
plt.savefig(r"C:\DDC\expenses_by_category.png")
Tables and charts from Revit in one click: a pipeline with CAD (BIM) data

The same idea as a video: tables and charts from a Revit project in one click. Step-by-step guides: install Jupyter Notebook and visualise data from Excel.

From people who automate

Why they work with pipelines more and more.

A few years ago Data-Driven Construction team showed me Jupyter Notebooks and I fell in love. Since then I do simple data manipulations quicker with a few lines of code instead of manually clicking in Excel. As always it was a gradual process, first I automated repeating tasks. As my confidence and know-how grows, I do more and more. Moreover, with the advent of Chat GPT, it became even easier, by describing the desired outcome I get the code snippets and just need to adapt them. It's saves me a lot of time. E.g. I have a list of new sign-ups and I have my company CRM. Once a month I check if all the signups are in the CRM. Before this was a manual comparison of two different lists. Now it's a press of a button and I get the list with the new signups in the right format to import it back to the CRM. By investing 1h of programming I'm saving 12h of work per year. I reduce failures dramatically. Because I did not like to compare two lists I did not do it as regularly as I should have done it. Moreover, I'm not good at comparing data, every phone call disrupted my work and I forgot to change a data entry. The consequences were that not everybody who should have gotten my newsletter got it. I feel like a god when I program. The more complex work becomes the less direct control over the results I have. But when programming I'm in absolute control of the result. I write a few lines of code and the computer executes them exactly as I tell him. When the result is not as I expected it to be, it's usually because of a logical shortcut on my side, the computer does exactly what I tell him.

Simon DilhasCEO & Co-Founder · Abstract AG

With the help of Python and especially the pandas library, as the DataDrivenConstruction team does, we are now able to perform delivery checks four times faster. By turning IFC data into a pandas DataFrame and comparing it with quantification tables or classification databases, we have made our expertise fast and reliable.

Dmitri GarbuzenkoAIM Coordinator · RB Rail AS

The decision to use Jupyter Notebook for results verification turned out to be highly beneficial. Our experience with solutions from Data Driven Construction and Jupyter Notebook has been extremely positive. Overall, the integration of solutions from Data Driven Construction and Jupyter Notebook has greatly improved our workflows. Their combined capabilities and performance have not only optimized data processing, but also produced meaningful results, making them highly recommended tools.

Irina FischerBIM Coordinator · OBERMEYER Group

Be part of the movement with DataDrivenConstruction! Let’s make true freedom in data formats a reality and catalyze a new era of productivity and innovation in construction. The real and necessary journey from closed to “open” formats has not even started: True freedom in data formats—free from proprietary ties and accessible to all—is what we need to unlock innovation and collaboration across the construction industry, particularly benefiting the 99% of construction, i.e. SMEs and unchaining the sector for accelerated growth

Prof. Dr.-Ing. Michael BühlerCo-Owner · GemeinWerk Ventures

Local automation

Everything runs on your machine.

No internet, cloud services or CAD/BIM licences required. Output is structured (SQL, Excel, HTML, JSON), so further analytics and automation are simple.

  • Complete data control
  • Security compliance
  • Independence from external services
  • Customisation for company-specific needs

Start simple

  1. 01Choose one routine taskFor example, extracting e-mail attachments or grouping Excel data.
  2. 02Use existing templatesThe public n8n library holds thousands of community workflows.
  3. 03IterateStart with a basic version and add functionality step by step.

Statistics and trends

A warning, and an opportunity.

3,292+

workflows in n8n's public marketplace in 2024.

n8n.io/workflows
9 to 12

months for n8n's user base to double, according to the company.

Highland Europe
Least digitised

Construction remains one of the least digitised industries in the global economy, according to McKinsey Global Institute.

Time to act

Get started with n8n, Dify, Flowise and LangFlow.

A new generation of tools makes integration and AI projects accessible to companies of any size, without a major investment in IT infrastructure. DataDrivenConstruction is an international team and community working on open construction data since 2021: we pick the first process, build the pipeline on your files and train your team.

Describe your process Automation workshop

Let’s talk

Let’s make your data work for you.

Tell us what slows your team down. You get an honest answer on what can be automated, how long it takes and how we would scope it. You pay only after the result.

  1. 01
    30-minute intro callYour data, your tools, the result you need. No sales script.
  2. 02
    Proof on your filesThen a written proposal with scope and timeline, agreed before any work starts.
  3. 03
    Pay after the resultYou pay when the agreed result is delivered.
[email protected]+49 152 58901584LinkedInTelegram

Greater Karlsruhe area, Germany · on-site in DACH, remote worldwide · EN / DE

Tell us about your project

Free intro call · NDA on request · no obligation

What do you need?

or e-mail [email protected]