Skip to main content
GET
Prefer the terminal? This endpoint is also a command in the Particle CLI: pick Particle CLI in the request example’s language menu. Install the Particle CLI.

Authorizations

​
X-API-Key
string
header
required

Pass your API key in the X-API-Key header (recommended).

Query Parameters

​
semantic_search
string

Vector-similarity search by meaning. Express the query the way you'd describe the topic to a colleague — paraphrase tolerant, finds segments that express the same idea in different vocabulary. Combine with keyword_search for hybrid ranking. Describe a topic, not a name: to find a specific person or company, filter with entity_id / company_id (or use /v1/podcasts/mentions for every line about them) — and for an exact token like a ticker, use keyword_search.

Maximum string length: 500
​
keyword_search
string

BM25 lexical search over dialogue. Use when an exact token must appear verbatim — tickers, drug names, product codes. By default every token must appear somewhere in the matched segment (see keyword_match). Wrap any token sequence in double quotes (e.g. "machine learning") to additionally require those words adjacent and in order in the segment's spoken dialogue — a speaker's own name is not indexed as dialogue, so quoting a person's name finds segments where the name appears in what was said, not segments they speak in. Adjacency spans consecutive lines, so a phrase may rarely straddle a line break. Multiple quoted phrases AND together (e.g. "central bank" "interest rates"). There is no boolean OR — a bare OR is treated as an ordinary word, so pass alternatives as separate requests.

Maximum string length: 500
​
keyword_match
enum<string>
default:required

How unquoted keyword_search terms are applied. 'required' (default) filters to segments containing every term, so results are guaranteed to contain what you typed — this also constrains semantic_search, making a hybrid query an intersection of the two. 'ranked' instead lets unquoted terms only boost BM25 relevance without excluding anything, which widens recall for loose bag-of-words queries but returns segments that may contain none of them. Quoted phrases are unaffected: they filter in BOTH modes, so to relax a phrase remove its quotes. Ignored when keyword_search is empty.

Available options:
required,
ranked
​
entity_id
string

Optional entity or person slug or ID. Filters results to episodes featuring this entity. A person with no linked entity matches the episodes they speak in. A slug that names both an entity and a person means the entity; pass the person's ID to filter by the person. For 'every line about X', use /v1/podcasts/mentions instead.

​
entity_type
string

Optional entity category slug. Filters to episodes that mention any entity in that category (e.g. 'company', 'school', 'book'). Speakers are always people, so this filter only narrows mentioned entities. Cannot be combined with 'role'. Ignored when entity_id resolves to a specific entity. See GET /v1/entities/types for the available slugs.

​
company_id
string

Optional company slug, domain, or ID. Resolves to the company's linked entity and applies as a filter.

​
speaker
string

Optional person slug or ID (a person's entity slug or ID works too). Keeps only segments in which this person speaks a line, and only their lines can be a window's match: 'did X say Y' rather than entity_id's 'Y came up in an episode with X'. With keyword_search, the person's own lines in the segment must carry every quoted phrase, and every unquoted term under keyword_match=required; a phrase may run across their consecutive lines but never into someone else's. Only speaking roles count: a clip of the person played on another show (a soundbite) is not them speaking, so their original turns up and its replays do not. With format=compact the page keeps the same shape and the same filtering, but loads transcripts to apply it. Retrieval narrows to the person's episodes before the line check, so lines of theirs that rank below other segments of those episodes can be missed: an empty page does not prove they never said it. Narrow with podcast_id, episode_id, since or until, or try semantic_search. Cannot be combined with the guest_ filters.

​
episode_id
string

Filter to a specific episode by slug or ID.

​
podcast_id
string

Podcast slug, internal ID, or numeric iTunes ID.

​
type
enum<string>

Segment type filter.

Available options:
INTRO,
PERSONAL_BANTER,
TOPIC_DISCUSSION,
INTERVIEW,
TRANSITION,
AD,
OUTRO
​
role
enum<string>

How the entity must relate to the episode. Speaking roles: 'guest', 'host', 'panelist', 'correspondent', or 'speaker' for any of them. 'mention' means the entity is talked about rather than speaking. Omit to match both. Matches the role values on /v1/podcasts/episodes.

Available options:
guest,
host,
panelist,
correspondent,
speaker,
mention
​
language
string

Restrict to episodes of podcasts in this language (ISO 639-1 code, e.g. 'fr'). Matches the podcast's primary language subtag, so 'fr' covers 'fr-FR'.

Example:

"fr"

​
since
string

Only segments from episodes published after this date.

Example:

"2024-01-01"

​
until
string

Only segments from episodes published on or before this date. A bare date includes that whole day.

Example:

"2024-06-01"

​
sort
enum<string>

Sort order. Defaults to relevance.

Available options:
relevance,
recency
​
format
enum<string>
default:full

Response shape. 'full' (default) carries each match's bounded dialogue windows and overlapping clips. 'compact' omits windows and clips and skips the transcript load and line scoring that produce them, so the page is a fraction of the size and returns sooner: each match keeps its episode, its segment (id, title, description, summary, bounds) and its relevance score. Use it to screen many results — a fan-out over companies, themes, or dates — and fetch dialogue only for the matches that survive, via GET /v1/podcasts/segments/{id}/transcript or a full-format search narrowed with episode_id.

Available options:
full,
compact
​
context
integer<int64>
default:1

Lines of surrounding dialogue around each matched line (1-15, default 1). Widens each match's transcript window in place — request more context instead of fetching the full transcript. Window size stays bounded even at the maximum. Ignored when format=compact.

Required range: 1 <= x <= 15
​
limit
integer<int64>
default:25

Results per page (1-100, default 25).

Required range: 1 <= x <= 100
​
cursor
string

Opaque cursor from a previous response.

​
guest_occupation
string

Only segments of episodes with a guest listed under this SOC 2018 or ISCO-08 occupation (code, slug, or title), including a SOC group or everyday name such as physicians or doctors. A word with no occupation filters by the field of research or capacity it names: oncologists is field 3211, founders is guest_capacity=executive_founder.

Maximum string length: 200
​
guest_field
string

Only segments of episodes with a guest listed under this ANZSRC 2020 field of research (code, slug, or title).

Maximum string length: 200
​
guest_standing
enum<string>

Only segments of episodes with a guest whose standing is recognized this way.

Available options:
established,
self_described,
unverified
​
guest_seniority
enum<string>

Only segments of episodes with a guest at this seniority or above.

Available options:
professional,
senior,
distinguished
​
guest_capacity
enum<string>

Only segments of episodes with a guest who knows their field in this capacity, such as executive_founder for founders. It narrows the other filters; an everyday word for a capacity passed as guest_occupation (founders, politicians) lands here too.

Available options:
researcher,
licensed_professional,
practitioner,
executive_founder,
public_official,
journalist_analyst,
creator_entertainer,
athlete_coach,
clergy,
firsthand,
layperson
​
guest_practicing
boolean

Only segments of episodes with a guest who still works in guest_occupation, a SOC 2018 occupation or group.

​
guest_in_field
boolean

Only segments of episodes with a guest speaking within their field (the guest the other guest_ filters select, when given).

​
speaker_statement_characteristics
object

Keep only lines said a certain way, by their speaker statement characteristics: speaker_statement_characteristics[certainty]=confident&speaker_statement_characteristics[temporal_orientation]=future finds confident predictions. Each key is an axis (sentiment, certainty, specificity, temporal_orientation) and its value one or more of that axis's labels, comma-separated: labels on one axis match any of them, and axes combine with AND. Every is_match line carries labels the filter accepts, and a segment is returned only when one of its lines does. Requires semantic_search, which names the topic: how something is said belongs here, never in the search text. With speaker or the guest_ filters, only the lines the selected people say count: confident predictions by cardiologists are guest_occupation=cardiologists with these. When more than 10,000 of their lines carry the labels, the newest 10,000 are searched; narrow with since, until, podcast_id or another label to reach older ones. The guest_ filters must name guest_occupation, guest_field or guest_capacity here and may select at most 5,000 guests with at most 50,000 appearances between them; narrow a broader filter with guest_practicing, guest_seniority or guest_standing. Cannot be combined with keyword_search alone, a quoted keyword_search phrase, or keyword_match=ranked.

Response

OK

​
data
object[]
required

Matches on this page, ranked by relevance (default) or recency

​
has_more
boolean
required

More results available

​
company
object

Resolved company, when company_id filtered the result set. Omitted when only entity_id was provided.

​
cursor
string

Opaque cursor; pass back as ?cursor= for the next page

​
diagnostics
object

Present only when data is empty: which filter excluded everything, why, and the parameter change to retry with.

​
entity
object

Resolved entity, when entity_id or company_id filtered the result set