Connect

POST the stream configuration and receive a node URL, a session id and a token.

Request

bash

curl -X POST https://api.lecsync.com/v1/realtime/connect \  -H "Authorization: Bearer lk_live_xxxxxxxx" \  -H "Content-Type: application/json" \  -d '{    "region": "us",    "audio":      { "encoding": "pcm_s16le", "sampleRate": 16000, "channels": 1 },    "transcribe": { "languages": ["en"], "strict": true, "speakers": false,                    "identifyLanguage": false, "displayMode": "sentence", "segmentation": "balanced" },    "translate":  { "languages": ["en", "zh"] },    "glossary":   { "terms": ["LecSync", "Soniox"],                    "pairs": [{ "source": "LecSync", "target": "录同步" }],                    "context": "University computer science lecture" },    "store": true  }'

Request fields

Field	Type	Description
`region`	`string?`	us \| uk \| jp \| ru. Omit it and we try the nearest region that still has capacity, using the request's source country. If that country is unknown, the order is us, uk, jp, ru. A specified region is never moved; an empty one returns 503.
`audio.encoding`	`string`	pcm_s16le — the only value in v1.
`audio.sampleRate`	`number`	16000.
`audio.channels`	`number`	1 (mono).
`transcribe`	`object`	Language hints and segmentation. See Transcription and Speakers.
`translate`	`object?`	Language pair. See Translation. Omit to disable translation.
`glossary`	`object?`	Terms and forced translations. See Glossary.
`aiEnhance`	`boolean?`	Second-pass refinement of each translation. See AI translation refinement.
`digest`	`boolean?`	Minutes streamed during the session. See Live minutes.
`store`	`boolean?`	Defaults to false (no audio or transcript persistence). Set true to retain audio + transcript for 30 days. See Recordings.

Response

json

{  "url": "wss://gw-us-a.lecsync.com/v1/sessions/cmxxxxxxxxxxxxxx",  "sessionId": "cmxxxxxxxxxxxxxx",  "token": "eyJhbGciOiJIUzI1NiJ9...",  "expiresAt": "2026-09-16T12:00:00Z",  "protocolVersion": 1,  "audio": {    "encoding": "pcm_s16le",    "sampleRate": 16000,    "channels": 1,    "maxFrameBytes": 65536  }}

Response fields

Field	Type	Description
`url`	`string`	WebSocket origin for this stream. Region-specific and per-stream — do not cache it.
`sessionId`	`string`	Identifies this stream, and the recording key when store is true.
`token`	`string`	One-time credential for opening the WebSocket.
`expiresAt`	`string`	ISO 8601 deadline for opening the connection, not a limit on its duration.
`protocolVersion`	`number`	1.
`audio`	`object`	The accepted audio format, including maxFrameBytes (65536).