---
title: CLI
description: Queria CLI commands, and output and authentication settings
---

## Commands

| Command | Purpose |
|---|---|
| `queria list` | List datasets |
| `queria search <keyword>` | Keyword search across datasets, tables, and columns |
| `queria info <dataset>` | Dataset metadata (license, source, etc.) |
| `queria schema <dataset>` | List a dataset's tables |
| `queria columns <dataset> [table]` | List columns (narrowed down when a table is given) |
| `queria summarize <table>` | Column statistics for a table (full scan) |
| `queria sql "<query>"` | Run SQL |
| `queria docs list` | List this documentation's pages |
| `queria docs show <name>` | Print one documentation page in the terminal |
| `queria skills list / show <name>` | List and print the published agent skills |
| `queria skills install [name]` | Install the agent skills into this project or machine |
| `queria login / logout` | Sign in with browser approval / delete the saved token |
| `queria auth set-token / status / clear` | Manage API tokens (relaxed rate limits) |
| `queria create / edit / delete` | Register one of your own datasets, change its visibility, or delete it |
| `queria telemetry status / disable / enable` | Manage anonymous usage telemetry |
| `queria dataset issue list / view / create / comment / close / reopen` | List, view, open, comment on, close and reopen a dataset's issues (wrong data or change requests) |
| `queria feedback send / status` | Send a request or a documentation problem |

## Output formats

Choose the standard output format with `--format`.

- `table` (default): a table laid out for reading in the terminal
- `csv`: CSV with a header row
- `json`: a JSON array of rows
- `jsonl`: one JSON record per line, for passing to other commands line by line
- `markdown`: a Markdown table, for pasting into docs and articles

```bash
queria list --format json | jq '.[].datasource'
```

In `table`, newlines inside a value become spaces. A value too wide for the terminal wraps onto more lines and is not cut off. Set a width with `COLUMNS=200 queria ...` to change where it wraps. Piped or redirected output is not wrapped and stays on one line per row.

## Save to a file

Add `--out` to save the results to a file. The extension determines the format (`.parquet` / `.csv` / `.json` / `.jsonl`). Save large results with `--out` rather than printing them.

```bash
queria sql "SELECT * FROM e_stat.ssds.a_pref_population" --out pop.parquet
```

## Automatic dataset ATTACH

When your `sql` names `<dataset>.<schema>.<table>`, the CLI attaches that dataset for you. When the SQL doesn't name the dataset, as when you read `information_schema`, name it with `--datasets`.

```bash
queria sql "SELECT * FROM information_schema.tables WHERE table_catalog = 'jma'" --datasets jma
```

When two accounts publish the same dataset name, write the handle as well: `--datasets alice/sales`. The name you use in SQL is still the short one (`sales.main.…`). To use both in one query, name them with `as`: `--datasets "alice/sales as a,bob/sales as b"`. To read your own private datasets, run `queria login` first or put a token in `QUERIA_TOKEN`.

## API tokens

Public data on `data.queria.io` can be read without a token. Reading without a token has rate limits, and registering a token raises them. You also need a token to read your own private datasets.

On your own computer, use `queria login`. When you approve in the browser, a token is issued and saved. Signing in again on the same computer replaces the token previously issued on that computer. Each computer you signed in on appears as "CLI (computer name)" under "Connected apps and devices" on the [API tokens](https://queria.io/settings/tokens) page, where "Disconnect" disables its token.

```bash
queria login
```

Where you can't approve in a browser, such as CI and BI tools, issue a token at [queria.io/settings/tokens](https://queria.io/settings/tokens) and pass it with `queria auth set-token <token>` or the `QUERIA_TOKEN` environment variable. See [Authentication](/en/authentication) for details.

## Feedback

From the terminal, `queria feedback send` sends dataset requests and documentation errors. For wrong data, open an issue on the dataset with `queria dataset issue create`. The full text of what will be sent is shown before it goes. See [Telemetry](/en/telemetry#sending-feedback) for details.

```bash
queria feedback send "I need childcare waiting-list counts by municipality" --kind dataset_request
```

Bug reports and feature requests are also welcome as [issues](https://github.com/queria-io/queria-cli/issues).
