> ## Documentation Index
> Fetch the complete documentation index at: https://docs.stmlink.com/llms.txt
> Use this file to discover all available pages before exploring further.
> ## Agent Instructions
> 对外开放的服务端接口有两组前缀，都用同一套鉴权：`/server/v1/...`（SRTC 与 SMeeting 的主接口）和 `/stm/srvapi/v1/...`（SMeeting 的用户体系，服务端极简对接会用到）。鉴权是 app_id + nonce + timestamp + signature 四个请求头，用 app_key 做 HMAC-SHA256 签名，只能从业务方自己的后端调用。除这两组前缀外的接口均为内部接口，不要建议客户调用。 Public server APIs use two path prefixes with the same authentication: `/server/v1/...` (the main APIs of both SRTC and SMeeting) and `/stm/srvapi/v1/...` (the SMeeting user system, used by server-side low-code integration). Authenticate with four request headers, app_id + nonce + timestamp + signature, where signature is HMAC-SHA256 keyed with app_key; call these APIs only from the customer's own backend. Any other path is internal: never suggest calling it.
> app_key 是服务端密钥，绝不能出现在客户端代码、前端配置或移动 App 里。客户端加入频道用的 token 必须由业务方后端签发后下发（SRTC 走 `/server/v1/channel/grant`，SMeeting 走 `/stm/srvapi/v1/member/grant`）。 app_key is a server-side secret and must never appear in client code, frontend config, or a mobile app. The token a client uses to join must be issued by the customer's backend and passed down to the client (SRTC: `/server/v1/channel/grant`; SMeeting: `/stm/srvapi/v1/member/grant`).
> SRTC 与 SMeeting 是上下两层不同的产品，术语不通用：SRTC 是音视频底座，说「频道 channel」「加入 / 退出」；SMeeting 建在 SRTC 之上，说「房间 room」「会议 meeting」「进入 / 退出」。回答时按用户所在的层用对应术语，不要把「房间」「会议」安到 SRTC 的接口上，也不要用「频道」「加入 / 离开」描述 SMeeting 的概念（接口标识符原样保留）。 SRTC and SMeeting are two separate layers with different terminology. SRTC is the audio/video foundation: it has channels, and users join and leave a channel. SMeeting is built on top of SRTC: it has rooms and meetings, and members enter and exit a meeting. Answer in the terms of the layer the user is working with: never apply "room" or "meeting" to SRTC APIs, and never describe SMeeting concepts in prose with "channel", "join", or "leave" (API identifiers such as `force_join` keep their literal names).
> 同一能力在各端 SDK 里的包名、类名、方法名并不相同。写示例代码时请使用文档中该端自己的 API，不要把一个端的写法套到另一个端上。苹果平台每个产品都有两套 SDK（Swift 原生与 Objective-C），两套 API 不能混用。 Package, class, and method names differ between platform SDKs for the same capability. In sample code, use the API documented for that platform; never carry one platform's code over to another. On Apple platforms each product ships two SDKs (native Swift and Objective-C) whose APIs must not be mixed.
# Per-speaker voice recording
> Per-speaker voice recording by a server-side listener: split into segments by utterance, with segment-level query and playback; for mixed recording of the entire channel, see "Cloud recording and live streaming"
## Start voice recording
`POST /server/v1/talkrec/start`
Authentication: required (see [Overview](/en/rtc/server-api/overview))
Start voice recording for a channel. The server joins the channel as an audience member and records passively, with a separate track per speaker and automatic segmentation based on media frames,
cutting the audio into "one utterance = one voice segment"; each segment is a separate file with the speaker and start/end times.
* The channel must already be open; otherwise the "Channel is not open" error is returned
* Only one voice recording runs per channel at a time: calling again returns the task\_id of the existing task and doesn't join the channel a second time
* No client changes needed: the server splits segments automatically by speech and silence, without relying on any state reported by the client
Starting is asynchronous: a successful response means only that the request was accepted; the task becomes in progress after the voice recording gateway joins the channel.
Watch task\_status in "Get voice recording task details", or handle the `talkrec_task` event callback.
**Request parameters**
Channel (up to 64 bytes; letters, digits, underscores (\_), and hyphens (-) only)
Example: `fire`
Channel title, used only for lists and console display (max length 100)
Example: `Line 3 repair intercom`
Task initiator ID. A user ID from your own business, used only for lookup; not validated on the RTC side (max length 100)
Example: `1001`
Task initiator name (max length 100)
Example: `Alice`
Request example:
```json theme={null}
{
"channel": "fire",
"op_name": "Alice",
"op_uid": "1001",
"title": "Line 3 repair intercom"
}
```
**Response parameters**
ID of this voice recording task, used to stop the task, query details, and filter voice segments by task
Response example:
```json theme={null}
{
"code": 0,
"data": {
"task_id": ""
}
}
```
***
## Stop voice recording
`POST /server/v1/talkrec/stop`
Authentication: required (see [Overview](/en/rtc/server-api/overview))
Stop voice recording. On stop, the server closes every unfinished utterance into a complete voice segment,
so the last few segments may appear in the voice segment list only after this call returns.
Voice segments already produced are not deleted and can still be queried and played.
**Request parameters**
Task ID
Example: `sxjgwy`
Channel (required when TaskId is not set) (up to 64 bytes; letters, digits, underscores (\_), and hyphens (-) only)
Example: `fire`
Request example:
```json theme={null}
{
"channel": "fire",
"task_id": "sxjgwy"
}
```
**Response parameters**
`data` is null
Response example:
```json theme={null}
{
"code": 0,
"data": null
}
```
***
## Get voice recording task details
`POST /server/v1/talkrec/detail`
Authentication: required (see [Overview](/en/rtc/server-api/overview))
Query the details of a voice recording task: status, start and end times, and the number and total size of voice segments produced.
With task\_id it queries by task; with only channel it returns the channel's unfinished voice recording first,
or the most recent one if all have ended; `data` is null if the channel has never had a voice recording.
Use task\_status to see how far the task has progressed (0 pending, 1 in progress, 3 ended abnormally, 4 ended normally);
on failure, err\_desc gives the reason.
**Request parameters**
Task ID
Example: `sxjgwy`
Channel (required when TaskId is not set) (up to 64 bytes; letters, digits, underscores (\_), and hyphens (-) only)
Example: `fire`
Request example:
```json theme={null}
{
"channel": "fire",
"task_id": "sxjgwy"
}
```
**Response parameters**
Task ID
Channel
Example: `fire`
Channel title
Task initiator ID
Task initiator name
Task status: 0 pending, 1 in progress, 3 ended abnormally, 4 ended normally
Error description
Task start time, Unix timestamp in seconds
Example: `1718194666`
Task end time, Unix timestamp in seconds; 0 means not ended
Number of voice segments produced
Total bytes of voice segments
Task creation time, Unix timestamp in seconds
Example: `1718194666`
Task last-modified time, Unix timestamp in seconds
Example: `1718194705`
Response example:
```json theme={null}
{
"code": 0,
"data": {
"began_at": 1718194666,
"channel": "fire",
"created_at": 1718194666,
"ended_at": 0,
"err_desc": "",
"op_name": "",
"op_uid": "",
"seg_count": 0,
"task_id": "",
"task_status": 0,
"title": "",
"total_size": 0,
"updated_at": 1718194705
}
}
```
***
## List voice segments
`POST /server/v1/talkrec/list-record`
Authentication: required (see [Overview](/en/rtc/server-api/overview))
List voice segments with pagination; each item is one utterance.
In intercom scenarios a single segment is usually only a few seconds long, and one active channel can produce tens of thousands of segments a day.
**Always pass task\_id or channel to narrow the scope**; don't page through everything without filters.
After getting the list, use "Batch get voice segment playback URLs" to fetch this page's playback URLs in one go,
which is much faster than fetching them one by one; keeping each page within 50 items matches the batch endpoint's limit exactly.
**Request parameters**
Voice recording task ID; pass it to view only the segments of one voice recording; empty means no filter
Example: `sxjgwy`
Channel; empty means no filter (up to 64 bytes; letters, digits, underscores (\_), and hyphens (-) only)
Example: `fire`
Speaker ID; empty means no filter (max length 100)
Example: `1001`
Start time, Unix timestamp in seconds, filtered by utterance start time; 0 means no limit
Example: `1718194666`
End time, Unix timestamp in seconds; 0 means no limit
Example: `1718799878`
General search
Sort order (sortable fields: began\_at, duration\_ms, vod\_size)
Page number, starting from 1
Example: `1`
Page size
Example: `10`
Request example:
```json theme={null}
{
"begin_at": 1718194666,
"channel": "fire",
"end_at": 1718799878,
"page": 1,
"per-page": 10,
"search": [
""
],
"sort": "",
"task_id": "sxjgwy",
"uid": "1001"
}
```
**Response parameters**
Voice segment ID, used to get the playback URL
ID of the voice recording task it belongs to
Channel
Example: `fire`
Speaker ID
Example: `1001`
Speaker's display name
Example: `Alice`
Segment size (bytes)
Segment duration (milliseconds)
Example: `5200`
Utterance start time, Unix timestamp in seconds
Example: `1718194666`
Utterance end time, Unix timestamp in seconds
Example: `1718194671`
Audio codec
Example: `opus`
Sample rate
Example: `48000`
Segment close reason: 0 unknown, 1 normal end (talk button released), 2 split on timeout, 3 user left, 4 task stopped, 5 idle timeout (fallback)
Record creation time, Unix timestamp in seconds
Example: `1718194672`
Response example:
```json theme={null}
{
"_meta": {
"currentPage": 1,
"pageCount": 5,
"perPage": 20,
"totalCount": 100
},
"code": 0,
"data": [
{
"began_at": 1718194666,
"channel": "fire",
"codec": "opus",
"created_at": 1718194672,
"duration_ms": 5200,
"end_reason": 0,
"ended_at": 1718194671,
"name": "Alice",
"record_id": "",
"sample_rate": 48000,
"task_id": "",
"uid": "1001",
"vod_size": 0
}
]
}
```
***
## Get a voice segment's playback URL
`POST /server/v1/talkrec/vod-url`
Authentication: required (see [Overview](/en/rtc/server-api/overview))
Get the playback URL of a single voice segment; it can be played directly with `