particle_podcast_get_bias_leaderboard

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

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.

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 (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. The per-podcast bias enum is already on every podcast result from particle_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):

## 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.

Example

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.
• For publishers concentrated in one bias bucket, use particle_podcast_list_bias_publishers.
• For one publisher’s full bias profile, use particle_podcast_get_publisher with include: ["bias"].