> ## Documentation Index
> Fetch the complete documentation index at: https://docs.particle.pro/llms.txt
> Use this file to discover all available pages before exploring further.

# particle_podcast_get_bias_leaderboard

> Rank podcast publishers by political-bias metrics across their analyzed catalogs.

<Note>
  In the **`podcast_bias`** category, which is **opt-in** — it is callable by name on any connection, but only advertised on `tools/list` when selected via `?include=podcast_bias`. See [Tool sets & discovery](/mcp/tool-sets).
</Note>

Rank podcast publishers by political-bias metrics across their analyzed catalogs: most left/right leaning (average lean score), most political (share of political shows), most diverse / most monolithic (lean spread), or most analyzed (coverage).

Publisher slugs feed [`particle_podcast_get_publisher`](/mcp/tools/podcast_publishers/podcast-get-publisher) (use `include: ["bias"]` there for one publisher's full bias profile). For publishers concentrated in one specific bias bucket, use [`particle_podcast_list_bias_publishers`](/mcp/tools/podcast_bias/podcast-list-bias-publishers). The per-podcast bias enum is already on every podcast result from [`particle_podcast_resolve`](/mcp/tools/podcasts/podcast-resolve).

## Inputs

| Field                    | Type            | Required | Default | Description                                                                                                                                                                                                                                         |
| ------------------------ | --------------- | -------- | ------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `metric`                 | string          | yes      | —       | Ranking metric. `most_left_leaning` / `most_right_leaning` rank by average lean score; `most_political` by share of political podcasts; `most_diverse` / `most_monolithic` by lean spread; `most_analyzed` by coverage.                             |
| `political_context`      | string          | no       | —       | Restrict the corpus to one political framework before aggregating. One of `US`, `UK`, `EU`, `CANADA`, `AUSTRALIA`, `INDIA`, `OTHER`, `UNKNOWN`.                                                                                                     |
| `min_analyzed_podcasts`  | integer (1–500) | no       | 5       | Only rank publishers with at least this many analyzed podcasts (aggregate signals are unreliable below it).                                                                                                                                         |
| `min_political_podcasts` | integer (1–500) | no       | 5       | Only rank publishers with at least this many political podcasts. Enforced only for the lean and diversity metrics (`most_left_leaning`, `most_right_leaning`, `most_diverse`, `most_monolithic`); ignored for `most_political` and `most_analyzed`. |
| `since`                  | string          | no       | —       | Only analyses evaluated on or after this ISO 8601 timestamp.                                                                                                                                                                                        |
| `until`                  | string          | no       | —       | Only analyses evaluated before this ISO 8601 timestamp.                                                                                                                                                                                             |
| `limit`                  | integer (1–50)  | no       | 10      | Publishers per page.                                                                                                                                                                                                                                |
| `cursor`                 | string          | no       | —       | Opaque pagination cursor.                                                                                                                                                                                                                           |

## Output

A markdown document with `## Publisher bias leaderboard — <metric> (N)` and a flat bullet list — one bullet per ranked publisher, formatted `#<rank> Publisher name (publisher-slug) — N analyzed, M political (P%), dominant <bias>, avg lean X.XX, lean stddev Y.YY` (the `political (P%)` share renders whenever the publisher has political podcasts; `dominant`/`avg lean` when a lean exists; `lean stddev` when the spread is defined — i.e. for the diversity metrics). When more pages exist, a horizontal rule and a `**Cursor:** <value>` line are appended.

Sample (`metric="most_right_leaning", limit=3`):

```markdown theme={"dark"}
## Publisher bias leaderboard — most_right_leaning (3)

- #1 The Daily Wire (the-daily-wire) — 11 analyzed, 11 political (100%), dominant RIGHT, avg lean 0.71
- #2 Blaze Media (blaze-media) — 9 analyzed, 9 political (100%), dominant RIGHT, avg lean 0.64
- #3 Salem Media (salem-media) — 28 analyzed, 24 political (86%), dominant LEANS_RIGHT, avg lean 0.55
```

The parenthesized `publisher-slug` feeds [`particle_podcast_get_publisher`](/mcp/tools/podcast_publishers/podcast-get-publisher).

## Example

```text theme={"dark"}
Agent calls: particle_podcast_get_bias_leaderboard {
  "metric": "most_left_leaning",
  "political_context": "US",
  "min_political_podcasts": 5,
  "limit": 25
}
```

## Related

* REST equivalent: [`GET /v1/podcasts/bias/publishers/leaderboard`](/podcasts/publishers-bias).
* For publishers concentrated in one bias bucket, use [`particle_podcast_list_bias_publishers`](/mcp/tools/podcast_bias/podcast-list-bias-publishers).
* For one publisher's full bias profile, use [`particle_podcast_get_publisher`](/mcp/tools/podcast_publishers/podcast-get-publisher) with `include: ["bias"]`.