Alternatives to Metaflow

Deepnote is a collaborative cloud notebook for Python and SQL where humans and AI agents build analyses, dashboards, and data apps together.

Open-source agentic AI data assistant that writes SQL, runs code in a sandbox, and chats with your self-hosted databases

Open-source SQL data warehouse on object storage — BI analytics, vector search, full-text search and geospatial in one engine

SQLBot turns plain-language questions into SQL, answers and charts over your own MySQL or PostgreSQL database — open source and self-hosted.

Daft is an open-source multimodal data engine for turning video, images, audio, and PDFs into training-ready datasets.

Open-source, MIT-licensed framework that runs LLM-driven generate-evaluate-rewrite loops for quantitative factor mining and ML experimentation.

AutoGluon: open-source AutoML that trains tabular, time series, and multimodal ML models in 3 lines of Python.

OneKE is an open-source bilingual knowledge extraction framework for NER, relation and event extraction from unstructured text.

Observable is a reactive JavaScript notebook for building interactive data visualizations and shareable data apps in the browser.

Python data apps with Plotly Dash: build interactive analytics dashboards using pure-Python callbacks and no front-end code.

Airbyte replicates data from 700+ sources and serves it to AI agents as governed context through its MCP server and Context Store.

Open-source GenBI engine that gives AI agents a governed context layer for SQL and dashboards

DataLab turns plain-English questions into runnable SQL, Python and R inside an AI data notebook you can review and share.

Kestra orchestrates data pipelines, infrastructure automation, and AI agent workflows from declarative YAML in one engine.

Livedocs is an AI data notebook where reactive cells re-run themselves and plain-English prompts write the SQL and Python.

TimeCopilot is an open-source AI forecasting agent that picks, validates, and explains time series models from a plain-English prompt.

Hex is an AI analytics workspace where a conversational agent and a SQL/Python notebook share one governed context layer.

Julius AI turns plain-English questions about your CSV, Excel, or Postgres data into charts, statistical tests, and regression models.

Databricks runs your lakehouse data, analytics and production AI agents on one governed platform across AWS, Azure and GCP.

OneSchema runs AI agents over 10GB+ files to ingest, normalize, and validate enterprise data for regulated industries.

Chat2DB turns plain English into runnable SQL across 40+ databases — with AI text-to-SQL that keeps your rows on your own machine.

Query any Hugging Face dataset as Parquet through a free REST API — no full download needed.

Healthcare price transparency data turns into negotiable rate benchmarks, provider rate scores, and traceable answers for US payer and

Granica is an efficiency layer for enterprise AI that cuts data lake storage costs and keeps long-running agent state alive.

GetDot is an AI data analyst that answers plain-English questions in Slack, Teams, or email with the number, the chart, and the SQL behind

Definite turns plain-English questions into self-hosted data apps backed by 500+ connectors and an AI analyst named Fi.

Jotform AI Data Assistant: build AI tables, charts, and bulk actions from your form submissions

marimo is a reactive Python notebook stored as Git-friendly .py files that runs as a script, app, or SQL workspace

Amazon SageMaker unifies data, analytics, and AI so teams can build, train, and deploy ML and foundation models on AWS.

Text2SQL.ai converts plain-English questions into dialect-correct SQL for 10+ databases, with a schema-aware assistant and a local-execution desktop app.