Reference

Run a sequence

Run up to 50 steps in order for ONE account: { op, input, continue_on_error } (any operation you may call, except dangerous ones), { wait_sec } or { wait_event: { type, timeout_sec } }. A failed / blocked step stops the rest unless continue_on_error. A later step may use an earlier result with "{{steps.N.result.field}}". Every step passes the same permissions and safety gate as a single call. Follow with sequences_get (wait ≤ 25 s).

View MarkdownThis page is generated from the source code on every release.

Overview

Name	Value
Code	`sequences.run`
Scope needed	`accounts.read`
Risk	Write
May need confirmation	No
Runs as	Answers right away
Needs `Idempotency-Key`	Yes
Member roles allowed	`owner`, `manager`, `agent`
In the app	Accounts → Log → Agent sequences

Input fields

Field	Type	Required
`accountId`	`uuid`	Yes
`title`	`text`	No
`steps`	`object[]`	Yes

Example request

curl -X POST https://tlg.builera.app/api/public/v2/ops/sequences.run \
  -H "Authorization: Bearer $TELEGRAM_AUTO_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: <unique key per operation>" \
  -d '{"accountId":"00000000-0000-0000-0000-000000000000","steps":[]}'

Output fields (`data`)

Field	Type	Required
`ok`	`true`	Yes
`sequence`	`object`	Yes

Example answer

{
  "data": {
    "ok": true,
    "sequence": {
      "id": "00000000-0000-4000-8000-000000000000",
      "title": "text",
      "account_id": "00000000-0000-4000-8000-000000000000",
      "connection_id": "00000000-0000-4000-8000-000000000000",
      "status": "running",
      "created_at": "2026-10-22T12:00:00.000Z",
      "finished_at": "2026-10-22T12:00:00.000Z",
      "steps": [
        "text"
      ]
    }
  }
}

Using it over MCP

Not listed on any profile: call tools_describe then tools_run with code: "sequences.run". Tool name when listed: sequences_run.

Run a sequence

#Overview

#Input fields

#Example request

#Output fields (data)

#Example answer