hm_timeline

The timeline is a chronological diary of what you and your agents did: what was stored, changed or deleted, plus entries written explicitly with hm_timeline_write. hm_timeline searches it by time, by node, or by text.

The timeline is never loaded automatically. Call it only when you need temporal context, so normal recall stays fast and focused on the graph.

Parameters

Parameter	Type	Default	Description
`query`	string	—	Case-insensitive text match on entries and their details.
`period`	string	—	A window ending now: `1h`, `3h`, `6h`, `12h`, `24h`, `7d`, `14d`, `30d`, `90d` or `1y`.
`start`	string	—	Start of a range, ISO 8601 (e.g. `2026-10-14T15:00:00`).
`end`	string	—	End of a range, ISO 8601.
`node_key`	string	—	Only entries about this node.
`limit`	integer	`50`	Maximum entries to return (up to 200).
`cursor`	string	—	The `next_cursor` from the previous page.

Response

Field	Description
`events`	Matching entries, newest first. Each has `id`, `ts`, `summary`, `source`, `tool`, `node_key` and `meta`.
`count`	Number of entries returned.
`next_cursor`	Cursor for the next page, or `null`.

Examples

{ "start": "2026-10-14T13:00:00", "end": "2026-10-14T18:00:00" }

{ "node_key": "decision_jwt_auth" }

{ "period": "1h" }

Equivalents

curl -X POST https://api.hypermemory.io/api/v1/memory/timeline \
  -H "Authorization: Bearer $HYPERMEMORY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query":"payment gateway","period":"30d"}'