Java and Python extensions I build for Altair AI Studio and RapidMiner — from an AI agent inside Studio to toolkits for documents, time series, and local models.
Filter by category. Each card shows what's inside, the package, and notes on the latest version.
An AI agent that lives inside Altair AI Studio. Chat with it from a dockable panel, connect it to the model provider you already use, and give it tools through MCP servers. It plans multi-step tasks, can look up examples in your own local repositories, and remembers corrections you teach it with /learn.
Chat panel · MCP client (stdio & HTTP/SSE) · OpenAI · Anthropic · Gemini · Azure OpenAI · Amazon Bedrock · Vertex AI · Ollama · DeepSeek · Databricks
A field experiment maintained by Marco A. Barradas — not an official Siemens product and not supported by RapidMiner. Conversation memory, learned lessons, and usage statistics are opt-in, stay on your computer, and are never transmitted.
Document processing for AI Studio pipelines: convert PDF, Word, Excel and PowerPoint files to Markdown or PDF, run OCR, extract text, tables and images, chunk content for RAG, compare document versions, and render Markdown back to HTML, PDF, DOCX or PPTX.
Convert File To Markdown · OCR Scan File · Extract Document File · Chunk Markdown File · Enrich Chunks Data · Compare Documents File · Crawl Site To Markdown · Markdown To PPTX File · and 31 more
Current build 0.9.167 (after 0.9.157 and 0.9.165). Conversion runs on MarkItDown or Docling, with Docling recommended for complex PDF layouts and tables.
Run Hugging Face models locally from AI Studio, without sending data to a cloud API. Covers text generation, embeddings and reranking, summarization, translation, NER and zero-shot classification, Whisper transcription, vision tasks from detection to segmentation, document QA, and vision-language models, plus dataset preparation and pushing to the Hub.
HF Local LLM · HF Local Embeddings · HF Reranker · HF Audio Transcription (Whisper) · HF Object Detection · HF Image Segmentation · HF Image-Text-to-Text (VLM) · HF Document QA · and 17 more
Version 2.0.0 targets the Python extension SDK and runtime 2.0. Imports are deferred and errors are handled row by row, so one bad input doesn't stop a batch.
End-to-end time-series work in AI Studio: explore seasonality, stationarity and autocorrelation, clean and transform series, then train, grid-search, backtest, compare and explain forecasting models built on Darts — and detect anomalies with PyOD models you can train, compare and evaluate.
Analyze Seasonality · Test Stationarity · Decompose Time Series · Train Forecast Model · Backtest Forecast Model · Compare Forecast Models · Train Ensemble Forecast · Detect Anomalies · and 21 more
Beta release 0.5.5 — every operator is marked BETA while the API settles. Forecasting uses Darts models; anomaly detection uses PyOD.
Tames messy spreadsheets: finds every table section on a sheet — stacked or side by side — stacks month-by-month tables into a single example set, pivots value columns by month, and splits wide data into equal attribute subsets.
Detect Table Sections · Append Month Tables · Append All Month Tables · Pivot Columns By Month · Split Attributes
Version 1.4.0 — five operators for sheet-level table wrangling, built on the Python extension SDK 1.1.
For Altair AI Studio / RapidMiner Studio 10.2 or later.
Request the .jar (Java) or .zip (Python) package for the extension you need.
Java: ~/.AltairRapidMiner/AI Studio/shared/extensions/ · Python: the python-extensions folder inside it.
The new operators appear in the Operators panel; AI Studio Agent adds its own chat panel.
Tell me what you're trying to build and we'll turn it into an AI Studio extension.
