Edit KB sync schedule
Set, change, or remove the auto-sync schedule on a URL or whole-site knowledge-base source, for example "re-fetch this page every day at 3am". Pass the source id (from list_kb_sources), a 5-field cron expression, and an IANA timezone; the next-run time is computed server-side. Each re-sync re-reads the source; pages whose content has not changed keep their passages, and changed pages are re-embedded (a metered charge). A failed re-sync keeps the source's previous passages. The schedule follows the source across drafts and publishes, and always refreshes the live copy. Set disable:true to remove the schedule. The sync does NOT run immediately; it fires on the schedule via the platform's cron tick. Time-based pull only; webhook push and CSS-selector targeting are not supported.
Authorization
bearerAuth A personal access token from Settings > API tokens.
In: header
Header Parameters
A key you choose, such as a UUID, that makes this POST safe to retry. Within 24 hours the same key and request answers with the first successful answer and runs nothing, the same key with a different request answers 422 idempotency_key_reused, and a key whose request is still running answers 409 idempotency_key_in_use. A failed request keeps nothing, so the same key runs it again.
1 <= length <= 255Request Body
application/json
TypeScript Definitions
Use the request body type in TypeScript.
Response Body
application/json
application/problem+json
application/problem+json
application/problem+json
application/problem+json
application/problem+json
application/problem+json
application/problem+json
curl -X POST "https://example.com/operations/edit_kb_source_schedule" \ -H "Content-Type: application/json" \ -d '{ "sourceId": "string", "agent": "string" }'{}{
"type": "http://example.com",
"title": "string",
"status": 0,
"detail": "string",
"code": "string",
"errors": [
{
"path": "string",
"message": "string"
}
],
"estimate": {},
"retryAfterSeconds": 0
}{
"type": "http://example.com",
"title": "string",
"status": 0,
"detail": "string",
"code": "string",
"errors": [
{
"path": "string",
"message": "string"
}
],
"estimate": {},
"retryAfterSeconds": 0
}{
"type": "http://example.com",
"title": "string",
"status": 0,
"detail": "string",
"code": "string",
"errors": [
{
"path": "string",
"message": "string"
}
],
"estimate": {},
"retryAfterSeconds": 0
}{
"type": "http://example.com",
"title": "string",
"status": 0,
"detail": "string",
"code": "string",
"errors": [
{
"path": "string",
"message": "string"
}
],
"estimate": {},
"retryAfterSeconds": 0
}{
"type": "http://example.com",
"title": "string",
"status": 0,
"detail": "string",
"code": "string",
"errors": [
{
"path": "string",
"message": "string"
}
],
"estimate": {},
"retryAfterSeconds": 0
}{
"type": "http://example.com",
"title": "string",
"status": 0,
"detail": "string",
"code": "string",
"errors": [
{
"path": "string",
"message": "string"
}
],
"estimate": {},
"retryAfterSeconds": 0
}{
"type": "http://example.com",
"title": "string",
"status": 0,
"detail": "string",
"code": "string",
"errors": [
{
"path": "string",
"message": "string"
}
],
"estimate": {},
"retryAfterSeconds": 0
}Dry run on a past call or chat POST
Try this version's hang-up and post-call functions on a past call or chat of this agent without sending anything. It builds the call's variables as the live run does (analysis results, post-call fields, caller details, pre-call outputs) and returns, for each function: its status (simulated means it would have run and sent, gated_off means its conditions did not pass, skipped or failed carries a reason), each condition rule with the value it read and whether it passed, and the rendered request or message (method, URL, headers and body, or the text, email or Slack message) with secret values masked. An app function also returns its action and its inputs converted to the types the action takes. A chat, or a call with no analysis, is analysed first with the same analysis the live run uses, including this version's post-call fields and AI-filled inputs; analysedNow is then true. That analysis is billed like a live one and is not saved, and only a member who can edit the agent can start it. A chat's hang-up functions see call status ended. When the agent keeps no transcripts, such a chat or call is refused, since there is nothing to analyse. Each row's aiUnfilled names the AI-filled inputs, and "message" for an AI-written text or body, that a saved analysis never wrote because it ran before they existed; those run empty. When contentKept is false the agent keeps basic call details only, so requests, responses and the values conditions read come back empty. Hang-up functions get no analysis results, as live. Nothing is written: no function log, no call record, no analysis and no field values. A tool function runs its tool in simulated mode, as in an eval. Without sessionId it uses the latest call or chat and also returns `calls`, the 20 most recent calls and chats, to choose from. Use when the operator asks to try, preview or dry-run the workflow, post-call or hang-up functions on a past call or chat.
Edit prompt POST
Make a TARGETED edit to the draft system prompt by a single find/replace span, far cheaper and faster than rewriting the whole prompt with `update_system_prompt`. PREFER THIS for small changes: a one-line tweak, tighten a rule, add a line to a section, fix a phrase, delete a stale sentence. `find` must match the current prompt exactly once (copy it verbatim, whitespace included); set `replace` to the new text (or empty to delete). Differs from `update_system_prompt`: that one full-replaces the entire prompt body, use it instead only when you're rewriting most of the prompt at once. The prompt guide still applies to the result (over the MCP, read_guide with 'system-prompt').