These automations and macros enhance mindmaps created by MindManager on macOS and Windows.
Using MindManager Macro:
Using Automator Workflow (Quick Action):
Using Tkinter UI and "Freetext" option (run the app_tkinter.py file):
More animated examples are in the doc folder.
- Azure OpenAI w/
GPTandomodels (use your key or log in withAzure Entra ID) - OpenAI w/
GPTandomodels (use your key) - Anthropic w/
Claudemodels (use your key) - xAI w/
grokmodels (use your key) - Google Gemini w/
Gemini,Gemmamodels (use your key) - Google Vertex AI w/
Geminimodels (use your access token / OAuth2) - DeepSeek w/
DeepSeekmodels (use your key) - Alibaba Cloud w/
Qwenmodels (use your key) - Mistral AI w/
Mistralmodels (use your key) - AWS Bedrock w/
Novanative models orAnthropic,Mistralserverless models (use your key + secret) - Azure AI Foundry (platform) w/
LLaMA,Mistraletc. (use your key) - Groq (platform) w/
OpenAI GPT-OSS,Gwen3,DeepSeek-R1,Kimi-K2,LLaMA3etc. (use your key) - STACKIT (Schwarz Gruppe) (platform) w/
LLaMA(use your key) - Fireworks AI (platform) w/
LLaMA,Qwenetc. (use your key) - Cerebras (platform) w/
GPT-OSS,Qwenetc. (use your key) - Perplexity (platform) w/
Sonarmodels etc. (use your key) - Hugging Face (platform) w/
LLaMAand more models (use your token) - GitHub Models (platform) w/ OpenAI
GPT,o, METALLaMAmodels etc. (use your key) - OpenRouter (platform) w/ many/all models (use your key)
- GPT4ALL (local w/ SDK) w/ any
llama.cppmodel - Ollama (local w/ API) w/ any
llama.cpporMLXmodel - LMStudio (local w/ API) w/ any
llama.cpporMLXmodel - MLX (local w/ API, Apple Silicon) w/ any
MLXmodel
- Azure OpenAI w/
gpt-image-1,DALL-E 3,Flux Pro 1.1,Flux.1 Kontext Pro(use your key or log in withAzure Entra ID) - OpenAI w/
gpt-image-1,DALL-E 3(use your key) - Stability AI w/
Stable Diffusion 3SD3.5/SD3/Ultra/Core(use your key) - Google Vertex AI w/
Imagen3(use your access token / OAuth2, GCP approval required!) - Ideogram AI w/
V1/V2(use your key) - Black Forest Labs w/
Flux Pro 1.1 Ultra,Flux Pro 1.1,Flux.1 Pro,Flux.1 Dev,Flux.1 Kontext Pro/Max(use your key) - Recraft AI w/
RecraftV3,Recraft20B(use your token) - Alibaba Cloud w/
Qwen Wan 2.2models - MLX (local w/ SDK, Apple Silicon) w/
Fluxmodels
- Azure OpenAI w/
Sora(use your key) - Google Vertex AI w/
Veo 3,Veo 2(use your access token / OAuth2)
- DeepL (use your key)
- Agno (formerly "Phidata")
- Autogen (Microsoft)
- CrewAI
- Windows compatible (run macro/context menu or call the Python script directly)
- macOS compatible (run Automator workflow (Quick Action) or call the Python script directly)
- Only native API requests to AI systems - no middleware needed
- Map format can be radial map or orgchart
- Using map templates on macOS
- Map styles on Windows are persistent, automatic collapsing of nodes
- Refinement of the map or topic.
- Refinement of the map or topic from a development perspective.
- Create examples for one, more (selected) or all topics.
- Clustering topics from scratch.
- Clustering by one or more criteria (e.g., Organization/Process/Project/Expertise, Capex–Opex perspective).
- Complex cases (multiple calls): eg. refinement + clustering + examples.
- Image generation from topics
- Professional translation of all topics by DeepL
- Export
Mermaidmindmap HTML document - Export
Markmapmindmap HTML document - PDF to mindmap (multiple files / batch processing)
- Generate a working paper (argumentation) HTML document for a detailed mindmap
- Generate a glossary HTML document of all terms
- Export mindmap to any other text format
- Change map layout by using a template (macOS)
- Reorder topics by business value or importance
- Misspelling or syntax correction
- Create a map based on external text data
There's a new user interface using Tkinter to execute most operations and more. The app stays always on top and has several UI tabs:
Choose the desired model and action and click on execute:
Choose the desired model and the number of images to generate:
Choose the desired model, enter text describing how to generate or modify the content, and click Execute.
Examples:
- "refine"
- "translate all topics to German"
- "add an emoji to every topic" (only works on macOS)
Choose the destination language and click Execute:
Choose an already implemented agent and the desired models.
The agents list is generated from scripts found in the ai/agents folder.
Before execution, the required libraries must be installed:
pip install -r requirements_agents.txt.
CrewAI does not install well on Windows ARM64 at the time of writing.
Agents make many AI round-trip calls, so costs should be monitored.
Currently, only the resulting chart format is selectable (orgchart, radial, or automatic):
First install the Windows package manager Chocolatey from an administrative shell or choose any other way following https://chocolatey.org/install.
This is one line:
Set-ExecutionPolicy Bypass -Scope Process -Force; [System.Net.ServicePointManager]::SecurityProtocol = [System.Net.ServicePointManager]::SecurityProtocol -bor 3072; iex ((New-Object System.Net.WebClient).DownloadString('https://community.chocolatey.org/install.ps1'))
Change to folder %localappdata%\Mindjet\MindManager\23\macros:
cd %localappdata%\Mindjet\MindManager\23\macros
Copy all files from the GitHub repository to this location.
Change to windows folder:
cd windows
Run install.bat or the following commands (requirements_auth.txt only if you want to use Azure Entra ID or GCP OAuth2):
cd ..
choco install python3
pip install -U -r .\src\requirements.txt
pip install -r .\src\requirements_auth.txt
powershell -ExecutionPolicy Bypass -File .\windows\macro_registration.ps1
Check in the registry and MindManager if the macros are available (right‑click on a topic).
Hint: The macro list is ordered according to the GUID string, not the macro name.
Macros can also be executed from the macro editor. The macros are similar except for the action parameter.
You can also check here if the path to the Python file is correct.
Python has to be installed first. Go to https://www.python.org/downloads/macos/ and download the desired installer.
Install required Python libraries (requirements_auth.txt only if you want to use Azure Entra ID or GCP OAuth2, requirements_mac_mlx.txt is only needed for local image generation using MLX):
pip install -U -r requirements.txt
pip install -r requirements_mac_mlx.txt
pip install -r requirements_auth.txt
Create the directory structure ~/git/mindmanager_ai with Terminal:
cd ~/
mkdir git
cd git
mkdir mindmanager_ai
cd mindmanager_ai
Copy all repository files to this location as the Automator workflows contain this path.
Alternatively, you can clone the repository in Terminal:
cd ~/
mkdir git
cd git
git clone https://github.com/robertZaufall/mindmanager_ai.git
cd mindmanager_ai
Change to folder macos and copy the Automator workflows to the ~/Library/Services (hidden) folder:
cd macos/automator
chmod +x ./copy_to_services.sh
./copy_to_services.sh
If you need elevated privileges for copying the files, use this command:
sudo sh ./copy_to_services.sh
All Automator workflow settings are similar but the action parameter:
The workflows are then available in the "MindManager" main menu -> Services
I prefer to execute the Python script directly from VSCode. Here you can easily adjust settings, try different LLMs on the fly, and debug if problems occur (external systems are sometimes not available).
Some actions are already predefined for quick execution.
There are main configuration files for LLMs, image generation, and translation.
Open each config file and uncomment the AI model you want to use. The config folder contains environment files for every supported AI provider. For example, if you want to use OpenAI models, copy config/openai.env.example to config/openai.env and fill in your API key.
Use the appropriate LLM system for which you have an API key. These keys are available on the developer platforms of the AI vendors.
If you want to run local models with Ollama, GPT4All, LMStudio, or MLX, you need either a newer Apple Mac model with an M1–M4 processor or a desktop/notebook with an NVIDIA graphics card with at least 8 GB of VRAM.
You can have more than one open document in MindManager. The document which should be processed must be the active document. For every processing a new document with the new topics will be created.
To process the whole map, select the central topic (for right‑clicking) or don't select any topic and call a macro manually (Windows), choose Automator Workflow from MindManager Menu -> Services, or call the Python script from VSCode or the command line: python3 process.py <action> <format>. If it doesn't work, try python3 or python.
Select the central topic or deselect all topics and call the automation.
You can also select one or more topics and start the automation for just these topics, e.g. to generate examples for these topics, refine just these topics etc.
Just select the topics for which you want to generate an image and choose the action "Generate Image" (macro on Windows or Automator Workflow on macOS) or call the Python script with parameter image or image_n.
After a while, the image will open and also be stored in the MindManager library Images folder.
Unfortunately, on macOS the image cannot automatically be inserted into the map or added to a topic due to insufficient library support.
On Windows the image can be automatically set as the background image of the map.
The results from the generation process are best with FLUX.1, good with DALL-E 3 and SD 3.5. Prompt crafting/engineering is still in progress.
The filename includes the generation seed where this feature is supported. This seed is useful if you want to generate similar images (e.g., with a different prompt). DALL-E 3 does not support a seed value anymore (at the time of writing).
The prompt for image generation can optionally be optimized using an LLM call.
Images can also be generated locally on macOS with Apple Silicon using the native Apple MLX framework.
Recently, more image‑generation platforms have emerged. Black Forest Labs, Ideogram AI, and Recraft AI image generation from mindmaps is already implemented and the results are excellent.
Video generation works with Sora (Azure OpenAI) and Veo (Google Vertex AI).
Put the files into the input-folder and use the action pdf_mindmap. The PDF files are first converted to markdown (MD) format. 'Reference' sections are removed as these contain no information but take a lot of tokens (e.g. arXiv papers). No OCR takes place by now. Tables are removed and the content will be highly sanitized by removing irrelevant characters, code blocks, href-links, whitespace etc.
Ollama‑based local LLMs are not working for me yet.
Some models (e.g., Sonnet) support native PDF processing, which is also implemented (action pdfsimple_mindmap).
Generating larger text outputs requires a model with a higher max‑token limit (e.g., GPT-4o, Gemini Flash, Sonnet). Results are very good most of the time.
The solution is best tested with Azure OpenAI. Results are perfect for most use cases.
Azure Entra ID authentication can be used in enterprise scenarios.
Gemini Pro results are best. Gemini Flash is also very good.
Vertex AI needs a security token, which you can generate using the Cloud Console.
The Sonnet model is very good, Haiku is good and also very cheap, Opus the most expensive.
Grok is very good and can refine mindmaps across several levels. The models grok-3 and grok-2-vision-1212 are very good. The vision model can be used for PDF OCR. grok-4 takes some time.
Amazon Bedrock has native models (e.g., Nova) and also hosts third‑party models from Anthropic Claude and Mistral.
DeepSeek created an extraordinary open‑source model, DeepSeek V3. The reasoning model r1 does not work yet.
Alibaba Cloud models cannot generate large amounts of tokens (Qwen-Max: 2000, Qwen-Plus + Qwen-Turbo: 1500) but the results are good. Qwen-Turbo is very fast.
Mistral AI hosts their commercial flagship models Mixtral-Large and Pixtral-Large. Mixtral-Large is a 'best in class' model. The maximum number of possible output tokens is a little unclear (max_tokens may be the sum of input and output tokens).
Groq is one of the fastest LLM hosting platforms. OpenAI GPT-OSS, Gwen3, DeepSeek-R1, Kimi-K2, LLaMA3 and other models are available. From time to time the supported models on the platform are changing.
Cerebras hosts OpenAI GPT-OSS, Gwen3, DeepSeek-R1, LLaMA3 and other models at high inference speeds. From time to time the supported models on the platform change.
Perplexity hosts its own Sonar models. From time to time the supported models on the platform are changing.
To access better models a pro subscription is needed. LLaMA-3-8B can still be used.
On the OpenRouter platform there are a variety of models and systems available. Fallback scenarios are supported. You also get access to the newest OpenAI models like o1-preview.
Results depend on the model used. LLaMA3, Zephyr, and Mixtral work well.
MLX results depend on the model used. LLaMA3 works well.
The solution is best tested with Azure OpenAI. Results are very good. There are issues with text generated in images. Azure Entra ID authentication can be used in enterprise scenarios.
Image generation with SD3.5 and SD3 is the most flexible, as you can use a seed value, negative prompt, etc. Prompt engineering is most important here, as the results are far from being perfect by now.
Image generation results are too simple by now as prompt engineering is also most important here. Imagen3 has the highest image resolution (1:1 with 1536x1536). Imagen3 is GA (globally available) but there is an approval process to get access to the API.
Image generation is quite good using the V2 model. When activating API access, keep in mind that generating an API key immediately results in a $40 bill.
Image generation is extraordinary. The flagship model is Flux Pro 1.1 / Ultra. As usual, tokens must be prepaid and you need to be accepted to access the platform.
Image generation is very good. There are many pre‑defined styles which can be activated as needed. Available models are RecraftV3 and Recraft20B.
This local image generation alternative is only available on macOS with Apple Silicon processors like M1 and higher. The results are above average using the Flux models.
Using the image_n action (e.g., image_10) generates a batch of images in a row. A pre‑execution step can optimize the prompt using an LLM call. If only one topic is selected a different prompt is used than when multiple topics are selected. Only the first‑level topics together with the central topic should be selected for better results.
When using this image‑generation approach, the desired model and embeddings tokenizer are downloaded automatically. The expected download size is up to 70 GB.
If you are downloading for the first time, you may have to log in at Hugging Face with your token first to agree to their terms of model usage: huggingface-cli login --token <xyz>. Downloaded models are cached at ~/.cache/huggingface.
Translation works for these languages:
# supported languages as source
# BG,CS,DA,DE,EL,EN,ES,ET,FI,FR,HU,ID,IT,JA,KO,LT,LV,NB,NL,PL,PT,RO,RU,SK,SL,SV,TR,UK,ZH
# supported languages as target
# BG,CS,DA,DE,EL,EN-GB,EN-US,ES,ET,FI,FR,HU,ID,IT,JA,KO,LT,LV,NB,NL,PL,PT-BR,PT-PT,RO,RU,SK,SL,SV,TR,UK,ZH
Source language will be detected automatically. Formality parameter is not supported for all languages, so it is disabled by now. Context parameter was not used as DeepL states it's deprecated.
API requests point to the free tier. If you have a paid subscription change the URL in the config.py.
There are only some example implementations using agentic frameworks. Results are requested as markdown code from the agents and are transformed to Mermaid mindmap by another LLM call. This should be addressed soon.
https://github.com/agno-agi/agno
Mighty but very performant and easy to use framework.
https://github.com/microsoft/autogen
Framework is complex and not so easy to use. There is also a fork with a similar name pyautogen, renamed as AG2.
https://github.com/crewAIInc/crewAI
Mighty and complex framework. Installation on Windows 11 on ARM requires VS2022 build tools. Installation on macOS seems to be easier but still has many dependencies.
Strategy:
The mindm component is used on both platforms for MindManager automation.
Link to GitHub Repository
Link to GitHub Pages
MindManager COM objects are addressed by using the PyWin32 library:
MindManager objects are addressed by using the AppScript library:
The Mermaid mindmap syntax is used when talking to the OpenAI LLM as an intermediate "language". Log file contents for input, output, prompt can be used in other use cases eg. mindmap visualizations in GitHub markdown files.
Log file contents:
Example using a Mermaid mindmap in a GitHub markdown file.
Code:
```mermaid
mindmap
Creating an AI Startup
Market Research
Identify Target Audience
Analyze Competitors
Understand Market Trends
Assess Market Needs
Evaluate Market Size
Business Model
Define Value Proposition
Choose Revenue Streams
Plan Monetization Strategy
Identify Cost Structure
Determine Key Partnerships```
GitHub rendering of the map:
mindmap
Creating an AI Startup
Market Research
Identify Target Audience
Analyze Competitors
Understand Market Trends
Assess Market Needs
Evaluate Market Size
Business Model
Define Value Proposition
Choose Revenue Streams
Plan Monetization Strategy
Identify Cost Structure
Determine Key Partnerships
You can also use the content inside the Mermaid online editor (https://mermaid.live/edit):
Remarks:
The Mermaid layer will be replaced by structured output using extended Mermaid syntax or JSON in the future.
Currently, this project is in the early development phase, and generated outputs may include errors. Automated testing has not yet been implemented.