https://api.particle.pro while writing this page. Paste them with your own key and you’ll see the same kind of results.
Get an API key
- Sign up or log in at the
- Create an organization and project
- Open the project’s API Keys section
- Click Create API Key and copy the key — it won’t be shown again
X-API-Key header. See API reference → Introduction for the alternate Authorization: Bearer form.
1. Find a person
Start with a name search./v1/entities?q=… matches against the knowledge graph and returns ranked results with a slug for each — the slug is the handle you’ll pass to every other entity-aware endpoint.
Response
sam-altman. Once you know an entity’s slug, you can also fetch it directly with GET /v1/entities/sam-altman to get the full record. Every endpoint that takes an {id} accepts the canonical ID or the slug — pick whichever you have. We’ll use sam-altman for the rest of the tutorial.
2. Read every line said about them
Jump straight from “this is the entity” to “here is every line of dialogue across the podcast catalog where they’re mentioned” — grouped by episode, recency-ordered, with the surrounding lines for context and the matched line flagged.Response (truncated)
windows — contiguous ranges of dialogue where the entity is mentioned, with context_lines of surrounding speech (default 2, configurable up to 20). Every line that actually names the entity has is_mention: true. The start_seconds and end_seconds deep-link into the audio.
Use
/v1/podcasts/mentions for “every line about an entity.” Use /v1/podcasts/episodes/search for “every segment that talks about a topic” (paraphrase-tolerant) or contains a specific phrase (BM25). They’re complementary — see Search for when to reach for which.entity_id with podcast_id to scope dialogue to a single podcast — for example, every line the All-In podcast has aired about OpenAI:
Response (truncated)
podcast_id for any other slug (acquired, lex-fridman, the-tim-ferriss-show) to ask the same question of a different audience.
3. Find episodes where they were actually a guest
Mentions tells you where someone was discussed. Episodes tells you where they were featured. Filter byentity_id plus a role to get just the episodes where Sam Altman was on the show as a guest — not the hundreds of episodes where his name comes up in passing.
Response (truncated)
role accepts guest, host, panelist, correspondent, or mention. Drop the filter and you’ll get every episode that touches the entity in any way — useful for broad recall, but for “where was Sam Altman actually interviewed” you want role=guest.
Hold onto the first result’s id (2cgrh56ehQYBlXwqecmZ9P above) for the next call.
4. Surface the highlight clips
Episode highlights are AI-extracted clips — typically a handful per episode — scored for engagement and classified by type (INSIGHTFUL, FUNNY, CONTROVERSIAL, AHA_MOMENT, NOTABLE_LINE, and more). Pull the clips from the episode you just found.
Response (truncated)
audio_url is a direct MP3 of the clip — embed it, drop it into a player, or attach it to a social post. intro_statement is a ready-to-use lead-in for show notes or share copy.
5. Search dialogue by topic
Mentions and episode filtering both pivot off who is being talked about. Search pivots off what is being said. Usesemantic_search for paraphrase-tolerant retrieval — describe the topic in your own words and the engine finds segments that discuss the same idea, even when they use different vocabulary.
Response (truncated)
semantic_search matched on meaning. Lines whose text drove the match are flagged is_match: true. The response also returns any highlight clips that overlap the segment in the same payload.
For exact-phrase matches (tickers, drug names, model numbers), pass keyword_search instead — or both, for hybrid retrieval. Add entity_id or company_id to scope a topical search to episodes featuring a specific person or company. The full surface is documented in Search.
6. Pivot to companies
Companies in Particle API are cross-referenced across every identifier system you might already have — SEC CIK, Wikidata QID, stock ticker, domain, and knowledge-graph slug. Look up Shopify by ticker:Response
id: 21IMlzahxqyE3wPovDXXM9):
/v1/companies?ticker=SHOP, ?cik=…, ?qid=…). The nested entity.slug (shopify) is the same kind of handle you used for Sam Altman in step 1 — feed it to any entity-aware endpoint to surface Shopify’s mentions, episodes, and clips across the catalog.
7. Get sponsor analytics in one call
Companies show up in podcasts in two ways: as topics of conversation and as advertisers. The per-company advertising endpoint gives you the sponsor view in a single request.Response (truncated)
What’s next
Concepts
IDs and slugs, pagination, errors, pricing weight, and choosing the right endpoint for the job.
Podcasts overview
Episodes, transcripts, segments, clips, advertising, and bias.
API reference
Every endpoint with request and response schemas.