tencent-meeting

Manage Tencent Meeting (腾讯会议) via REST API. Schedule/create meetings, query meeting details, list cloud recordings, and extract meeting transcripts. Use when...

Manage Tencent Meeting via the open REST API (AK/SK auth).

Setup

Required env vars (set in shell or .env):

Var	Description
TM_SECRET_ID	API SecretId from Tencent Meeting open platform
TM_SECRET_KEY	API SecretKey
TM_APP_ID	Enterprise AppId
TM_SDK_ID	Application SdkId
TM_STS_TOKEN	(Optional) STS ticket for transcript/record APIs

Get credentials at: https://meeting.tencent.com/open-api.html

Commands

Create a Meeting

node {baseDir}/scripts/create_meeting.js '<JSON>'

JSON fields:

Field	Required	Description
userid	yes	Creator's user ID
subject	yes	Meeting subject
start_time	yes	ISO 8601 datetime or unix timestamp (seconds)
end_time	yes	ISO 8601 datetime or unix timestamp (seconds)
type	no	0=scheduled (default), 1=instant
password	no	4-6 digit meeting password
instanceid	no	Device type, default 1 (PC)
invitees	no	Array of {userid} objects
settings	no	Meeting settings (mute, waiting room, auto-record, etc.)
time_zone	no	Timezone string
location	no	Meeting location (max 18 Chinese chars)

Example — schedule a meeting tomorrow at 2pm:

node {baseDir}/scripts/create_meeting.js '{"userid":"user123","subject":"Weekly Sync","start_time":"2026-03-11T14:00:00+08:00","end_time":"2026-03-11T15:00:00+08:00","settings":{"auto_asr":true}}'

Query a Meeting

node {baseDir}/scripts/query_meetings.js --meeting-id <ID> --userid <UID>
node {baseDir}/scripts/query_meetings.js --meeting-code <CODE> --userid <UID>

List Cloud Recordings

node {baseDir}/scripts/list_records.js --meeting-id <MID> --userid <UID>

Get Record Detail (download URLs, AI summary)

node {baseDir}/scripts/list_records.js --meeting-record-id <MRID> --userid <UID>

Extract Meeting Transcript

# Text format (readable)
node {baseDir}/scripts/get_transcript.js --record-file-id <RFID> --operator-id <UID> [--meeting-id <MID>] [--format text]

# JSON format (full detail with timestamps)
node {baseDir}/scripts/get_transcript.js --record-file-id <RFID> --operator-id <UID> --format json

# AI-optimized transcript
node {baseDir}/scripts/get_transcript.js --record-file-id <RFID> --operator-id <UID> --type 1

Output (text format):

[00:04] Speaker1: 看一下那个哪能看。
[00:12] Speaker1: 可以。
[00:43] Speaker2: 没有类型还有...

Workflow: Get Transcript from a Meeting

1. Query the meeting to get meeting_id
2. List recordings: list_records.js --meeting-id <MID> --userid <UID>
3. From the response, get record_file_id
4. Extract transcript: get_transcript.js --record-file-id <RFID> --operator-id <UID>

Natural Language Mapping

User says	Action
"Schedule a meeting at 3pm tomorrow"	create_meeting.js with computed start/end times
"Create a meeting with password"	create_meeting.js with password field
"Get the transcript for meeting X"	Workflow: query → list records → get transcript
"Show meeting details"	query_meetings.js
"Download meeting recording"	list_records.js → use download URLs

API Reference

For detailed API fields and parameters, see references/api-guide.md.