Track a whole Discord server as a keyword
Creates a keyword for the connected server (kind: discord_server, named after the server) and binds every message of the server (or of channelIds) to it, without a text match; each message is still checked for relevance. It uses one ordinary keyword slot (400 KEYWORD_LIMIT_EXCEEDED when none is left) and pauses, resumes and deletes like any keyword (/api/v2/keywords/{id}). Idempotent per server: when the workspace already tracks it, the existing keyword and binding are answered with alreadyExisted: true and nothing is created, even at the keyword limit. Its posts count toward the workspace’s monthly mention allowance; dailyCap (1–2000) adds an optional per-UTC-day budget, and omitting it or sending null means none. Every channelIds entry must be a readable channel of the server inside its channel scope (400 DISCORD_CHANNEL_INVALID). Ordinary keywords need none of this: they match Discord messages when their platforms include discord. Requires an admin-scoped key; 404 DISCORD_LINK_NOT_FOUND for a disconnected or another workspace’s server, 403 DISCORD_SOURCE_UNAVAILABLE while the Discord source is not enabled for the workspace.
Authorizations
Clerk API key. Create one in Settings → API Keys. Pass as Authorization: Bearer <key>. Keys carry a scope — read, write (implies read) or admin (implies write) — chosen when the key is minted. Each operation's security requirement (and its x-required-scope extension) names the minimum scope it needs; request the least-privileged key that covers the operations you call.
Path Parameters
Body
Channels to collect; omit or [] for every channel the server link reads. Each id must be a readable channel of the server and inside its channel scope.
500A Discord snowflake id as a decimal string (64-bit — never parse it as a number).
^\d{1,20}$Response
200 response
