Skip to content

Search

Search runs across two kinds of results: knowledge topics and chapters of source content (sections of a document or transcript). It combines keyword and semantic matching, reranks the results, and returns them ordered by relevance, scoped to what you can see. Search always uses the latest state of the knowledge base.

GET /v1/search

ParameterTypeNotes
qstringThe search query. Required.
typesstringComma-separated result types: topic, chapter. Defaults to all.
limitnumber1–50. Default 10.
Terminal window
curl "https://api.mereon.ai/v1/search?q=expense+approval&limit=5" \
-H "Authorization: Bearer $MEREON_TOKEN"
{
"data": [
{
"type": "topic",
"id": "top_9f2a...",
"title": "Expense approval",
"snippet": "Expenses over $500 require manager sign-off...",
"score": 0.91,
"url": "https://mereon.ai/topics/expense-approval"
},
{
"type": "chapter",
"id": "chp_1b7c...",
"title": "Submitting an expense report",
"snippet": "...attach receipts and submit before month end...",
"score": 0.78,
"url": "https://mereon.ai/content/finance-handbook#c3",
"contentId": "cnt_44de...",
"contentTitle": "Finance Handbook",
"contentType": "document"
}
]
}
FieldTypeNotes
typestringtopic or chapter.
idstringTopic id or chapter id, depending on type.
titlestringResult title.
snippetstring | nullShort excerpt of the matched text, when available.
scorenumberRelevance score. Opaque: meaningful only for ordering within one response.
urlstringDeep link into Mereon.
contentIdstringchapter only. The content the chapter belongs to.
contentTitlestringchapter only.
contentTypestringchapter only, for example document or video.

POST /v1/search

Identical to the GET form, but accepts the same parameters as a JSON body. Use it for long queries or agent payloads. This POST is still a read; it changes nothing.

Terminal window
curl -X POST https://api.mereon.ai/v1/search \
-H "Authorization: Bearer $MEREON_TOKEN" \
-H "Content-Type: application/json" \
-d '{ "q": "how do we approve expenses over $500", "limit": 5 }'