GET returns the matching messages in the same response.
Search
server_id or channel_id. List the servers and channels available to filter on with retrieve_discord_servers — set include_channels=true to expand each server’s channels:
retrieve_discord_search_history:
Request parameters
create_discord_search takes these query parameters:
retrieve_discord_servers takes skip (min 0, default 0), limit (1–1000, default 100), and include_channels (boolean, default false). retrieve_discord_search_history takes page (min 1, default 1) and page_size (min 1, default 20).
Each search costs 0 credits. Rate limit: 1,000 requests/day, applied per endpoint.
What comes back
The search response is an object withresults, total, and search_history_id. Each item in results carries the message and its context.
Top-level fields:
results— array of matching messages.total— total number of matches for the query.search_history_id— id of this search in your history, ornullif it was not saved.
content— the full message text.content_highlighted— the same text with the matched term marked up.text_preview— a short preview of the message.posted_at— when the message was posted.edited_at— when it was last edited, ornull.is_pinned— whether the message is pinned in its channel.author— the poster, as{ name, id }.channel— the channel it appeared in, as{ name, id }.server— the server it belongs to, as{ name, id }.context— the surrounding text, as{ before, match, after, match_position }.embeds— array of embed objects attached to the message.
retrieve_discord_servers returns total, skip, limit, and a servers array. Each server carries id, name, description, member_count, icon_url, discovered_at, and last_scraped_at; with include_channels=true it also carries a channels array, where each channel has id, name, type, and message_count. retrieve_discord_search_history returns total, page, page_size, and an items array, where each item has id, query, searched_at, and total_matches.
When a hit lands, read the context to confirm the term is a real mention and not a coincidence, then pivot on author.id and server_id to map the actor and the community before tasking deeper collection.