Glossary

Key terms used throughout the Determ API documentation.

A

Access Token

A static Bearer token issued per user that authenticates every API request. Each user in an organization on an API plan has their own token. Pass it in the Authorization: Bearer YOUR_API_TOKEN header.

Author

The name or handle of the person or account who created the mention (e.g. a journalist byline, a tweet author, a Reddit username). Keywords can include or exclude specific authors; see Filtering by Author.

Auto Sentiment

The automatically detected sentiment of a mention. Returned as POSITIVE, NEUTRAL, NEGATIVE, or UNDEFINED. Different from manual sentiment labels that users can apply.

B

Bearer Token

Standard HTTP authentication scheme where a token is sent in the Authorization header prefixed with Bearer . See Access Token.

E

Engagement

A composite metric that reflects audience interaction with a mention relative to the source's typical reach. Higher engagement means the mention generated unusually high activity for its source. Available as a sort property in scroll endpoints.

Engagement Rate

The ratio of Interactions to Reach for a mention, expressed as a percentage or normalized score. Not currently returned by any documented endpoint — it exists in the Determ platform but is not part of the public API yet.

F

Feed

A stream of mentions collected for a specific keyword or group. Each feed has a feedTime (when the mention entered the feed) and a publishedTime (when the mention was originally published).

Feed Time

The timestamp when a mention was ingested into the Determ system. Different from Published Time. Use feedTime filters when you care about when data became available in your account; use publishedTime when you care about when content was actually written or posted.

FilterClause

The pattern used for mention filters in the API. Each filter supports any (match any of these values), all (match all), must (required), and not (exclude) sub-fields. Example: { "any": ["POSITIVE", "NEUTRAL"] }.

G

Group

A folder that organizes keywords within an organization. Groups help structure monitoring by theme, brand, competitor, or campaign. Each group contains one or more keywords.

I

Influence Score

A numeric score (0–10) representing the authority and credibility of a mention's source. Higher scores indicate more influential sources such as major news outlets or high-follower social accounts.

Influencer

A named entity (usually a journalist, public figure, or high-reach social account) associated with a mention. Distinct from author: an article written by staff reporter "John Doe" may be associated with an influencer brand like "The New York Times".

Interactions

The total engagement count for a mention — likes, shares, comments, retweets, reactions combined into a single number.

K

Keyword

A monitoring query that defines what content Determ tracks. A keyword includes a search query (natural language or boolean), language filters, source type filters, and other criteria. Keywords belong to a group within an organization.

M

Mention

A single piece of content — an article, social media post, forum comment, TV segment, radio clip, or print article — that matches a keyword query. Each mention has metadata including source type, sentiment, reach, author, and timestamps.

Mention Type (Source Type)

The channel where a mention was found. Supported types: WEB, FACEBOOK, TWITTER, YOUTUBE, INSTAGRAM, REDDIT, TIKTOK, FORUM, COMMENT, PRINT, RADIO, TV, BLUESKY.

N

Natural Query

A search query expressed in natural language with boolean operators. Example: Tesla OR "Elon Musk" NOT SpaceX. Supports AND, OR, NOT, phrase matching with quotes, and parentheses for grouping.

Natural Filters

Keyword filters written as plain text, one key:value1,value2 filter per line, e.g. author:"Starbucks Coffee". Passed in the naturalFilters field of the keyword object; see Filtering by Author.

O

Organization

The top-level account entity in Determ. An organization contains groups, keywords, tags, and users. All API requests are scoped to an organization via organizationId.

P

Problem Details (RFC 7807)

The standard format used for all error responses. Returns JSON with type, title, status, detail, instance fields, plus an optional code property on domain errors. Content type is application/problem+json. See Errors & Rate Limits.

Pagination

The mechanism for retrieving large result sets in chunks. The Determ API uses token-based pagination via Scroll Token on scroll endpoints. Each response includes a token to fetch the next page. A single page contains at most 1000 mentions (paged.count ≤ 1000).

Published Time

The timestamp when content was originally published at its source (e.g. when an article was published, when a tweet was posted). Different from Feed Time.

R

Rate Limit

A per-user quota on how many requests can be made to each endpoint within a time window. Enforced via token bucket algorithm. When exceeded, the API returns HTTP 429. See Errors & Rate Limits for per-endpoint limits.

Reach

The estimated audience size that was exposed to a mention. For web articles, this is typically the site's monthly visitors. For social media, it's the author's follower count. Measured as an integer.

Relative Time Period

A preset time range used in query filters. Values: TODAY, YESTERDAY, LAST_24_HOURS, LAST_7_DAYS, LAST_30_DAYS, LAST_90_DAYS, THIS_MONTH, THIS_WEEK, LAST_WEEK, LAST_MONTH. Use instead of specific from/to dates when you want a rolling window.

Rule

In the Determ app, the matching criteria attached to an automated action. A rule is not a Keyword, and this API exposes no endpoints for rules. If you came here looking to stop monitoring a topic, that is a keyword — see Delete Keyword.

S

Scroll / Scroll Endpoint

An endpoint that returns paginated mention data. Uses token-based pagination — the first request returns a Scroll Token that must be passed in subsequent requests to get the next page. Each page returns at most 1000 mentions; scrolling exists so you page through large result sets instead of fetching everything in one request.

Scroll Token

An opaque pagination token returned by scroll endpoints. Pass it in the next request to retrieve the following page of results. When null or absent, there are no more results.

Sentiment

The automatically detected tone of a mention. Values: POSITIVE, NEUTRAL, NEGATIVE, or UNDEFINED (when sentiment could not be determined).

Sort Property

The field used to order mention results in scroll endpoints. Supported values: PUBLISHED_TIME, FEED_TIME, REACH, VIRALITY, INTERACTIONS, INFLUENCE, ENGAGEMENT.

Source Reach

The reach value of a mention's source (e.g. follower count for a Twitter profile, monthly visitors for a website) rather than the specific piece of content. Not currently returned by any documented endpoint — mentions expose Reach instead.

T

Tag

A label that can be applied to mentions for manual categorization. Tags are organization-scoped and help teams collaborate on mention triage and analysis.

Token Bucket

The rate limiting algorithm used by the Determ API (Bucket4j implementation). Each endpoint has a bucket with a fixed capacity that refills over time. Each request consumes one token; when empty, requests are rejected with HTTP 429.

V

Virality

A score (0.0–1.0) indicating how rapidly a mention is being shared and discussed. Higher values indicate content that is spreading quickly across channels.