# undatum Command-line tool for converting, validating, querying, and documenting datasets across 140+ formats (iterabledata), with a streaming-first CLI. Entry points: `undatum` and `data`. Site: https://datenoio.github.io/undatum/ CLI help: `undatum --help` ## Getting started - [Installation](https://datenoio.github.io/undatum/getting-started/installation) — extras, uv/pipx, Python 3.9–3.13 CI, binaries - [Quick start](https://datenoio.github.io/undatum/getting-started/quick-start) - [Cookbook](https://datenoio.github.io/undatum/getting-started/cookbook) - [Basic usage](https://datenoio.github.io/undatum/getting-started/basic-usage) — `--filter` syntax, encoding, compression - [Shared CLI options](https://datenoio.github.io/undatum/commands/shared-options) — `--flatten-nested`, `--on-error`, `--error-log`, `--table`, `--quotechar` - [Performance](https://datenoio.github.io/undatum/getting-started/performance) - [Troubleshooting](https://datenoio.github.io/undatum/getting-started/troubleshooting) ## Commands - [Command index](https://datenoio.github.io/undatum/commands/) - [convert](https://datenoio.github.io/undatum/commands/convert) - [repack](https://datenoio.github.io/undatum/commands/repack) - [formats](https://datenoio.github.io/undatum/commands/formats) - [analyze](https://datenoio.github.io/undatum/commands/analyze) - [stats](https://datenoio.github.io/undatum/commands/stats) (`profile` alias) - [select](https://datenoio.github.io/undatum/commands/select) — `--filter` (alias `--filter-expr`) - [validate](https://datenoio.github.io/undatum/commands/validate) - [schema](https://datenoio.github.io/undatum/commands/schema) - [sql](https://datenoio.github.io/undatum/commands/sql) - [pipeline](https://datenoio.github.io/undatum/commands/pipeline) — YAML DSL, templates including `jsonl-normalization` - [apply](https://datenoio.github.io/undatum/commands/apply) — `--script` or `--plugin` - [db](https://datenoio.github.io/undatum/commands/db) — query, load, dump - [ingest](https://datenoio.github.io/undatum/commands/ingest) - [api](https://datenoio.github.io/undatum/commands/api) - [ai](https://datenoio.github.io/undatum/commands/ai) — `ai plan SOURCE TARGET` (two paths, not `--to`) - [doc](https://datenoio.github.io/undatum/commands/doc) - [mcp](https://datenoio.github.io/undatum/commands/mcp) - [tui](https://datenoio.github.io/undatum/commands/tui) / [web](https://datenoio.github.io/undatum/commands/web) - [config](https://datenoio.github.io/undatum/commands/config) - [plugins](https://datenoio.github.io/undatum/commands/plugins) - [extract](https://datenoio.github.io/undatum/commands/extract) — `undatum[extract]` - [package](https://datenoio.github.io/undatum/commands/package) ## Integrations - [MCP server and tool catalog](https://datenoio.github.io/undatum/integrations/mcp) - [Python SDK](https://datenoio.github.io/undatum/integrations/sdk) — Dataset.read/write, fill, dedup, sort, filter, select, join, sample, mask, rename, explode, enum, reverse, stats, package, convert_many - [Data API](https://datenoio.github.io/undatum/integrations/data-api) — optional `--api-key` / `UNDATUM_API_KEY` - [Cloud storage](https://datenoio.github.io/undatum/integrations/cloud) — `s3://`, `gs://`, `az://` - [Plugins](https://datenoio.github.io/undatum/integrations/plugins) - [AI documentation](https://datenoio.github.io/undatum/integrations/ai) ## Formats - [Format support matrix](https://datenoio.github.io/undatum/formats/) - [iterabledata formats](https://datenoio.github.io/iterabledata/formats/)