> ## Documentation Index > Fetch the complete documentation index at: https://docs.glot.com/llms.txt > Use this file to discover all available pages before exploring further. # Create a translation room > Create a live translation room. **Languages are not set here.** Each participant declares their own when you request their join token, so you can create a room before knowing who will be in it or what they speak. The translator joins on its own once the room holds two languages, and leaves when fewer than two remain — you never dispatch or dismiss it yourself. The response carries the room's `id`, for looking up its usage later, and the realtime `url` your clients connect to. Get a token for each participant, including the first, from `POST /v1/rooms/{room_name}/token`. `room_name` must be unique among rooms that are currently live; reusing the name of a live room returns 409. Omit it to have one generated. `streaming_mode` records how the room's audio will arrive — `mono` when both languages come from one device, `dual` when each speaker has their own. It is optional: omit it and the room's **first join token** decides, by how many languages it asks for. Either way the mode is fixed for the room's lifetime and every later join token must match it. Requires credit: a room cannot be started while the organization's balance is at or below zero, and the request returns 402 if it is. A call already in progress is never cut off, so a long room can run the balance negative. ## OpenAPI ````yaml https://api.staging.glot.com/openapi.json post /v1/rooms openapi: 3.1.0 info: title: glot-api version: 0.1.0 servers: [] security: [] tags: - name: rooms description: Create translation rooms and let participants join them. - name: usage description: Minutes and charges over time, for reporting and charts. - name: billing description: Buy prepaid credit and review the balance behind it. - name: api-keys description: Create and revoke keys used to authenticate requests. - name: supported-languages description: Languages available for translation. - name: users description: The authenticated user's own profile. paths: /v1/rooms: post: tags: - rooms summary: Create a translation room description: >- Create a live translation room. **Languages are not set here.** Each participant declares their own when you request their join token, so you can create a room before knowing who will be in it or what they speak. The translator joins on its own once the room holds two languages, and leaves when fewer than two remain — you never dispatch or dismiss it yourself. The response carries the room's `id`, for looking up its usage later, and the realtime `url` your clients connect to. Get a token for each participant, including the first, from `POST /v1/rooms/{room_name}/token`. `room_name` must be unique among rooms that are currently live; reusing the name of a live room returns 409. Omit it to have one generated. `streaming_mode` records how the room's audio will arrive — `mono` when both languages come from one device, `dual` when each speaker has their own. It is optional: omit it and the room's **first join token** decides, by how many languages it asks for. Either way the mode is fixed for the room's lifetime and every later join token must match it. Requires credit: a room cannot be started while the organization's balance is at or below zero, and the request returns 402 if it is. A call already in progress is never cut off, so a long room can run the balance negative. operationId: create_room_v1_rooms_post requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/CreateRoomRequest' responses: '201': description: Successful Response content: application/json: schema: $ref: '#/components/schemas/CreateRoomResponse' '422': description: Validation Error content: application/json: schema: $ref: '#/components/schemas/HTTPValidationError' security: - HTTPBearer: [] components: schemas: CreateRoomRequest: properties: room_name: anyOf: - type: string - type: 'null' title: Room Name description: >- Optional. A name for the room, which must not be in use by another room that is currently live. **Omit it and a unique name is generated**, returned as `room` in the response. examples: - standup streaming_mode: anyOf: - $ref: '#/components/schemas/StreamingMode' - type: 'null' description: >- Optional. How this room's audio will reach the translator: - `mono` — both languages come from the same device, such as two colleagues sharing one phone. - `dual` — each speaker joins from their own device. **Omit it and the first join token decides**: a token requesting two languages makes the room `mono`, one language makes it `dual`. Either way the mode is fixed for the room's lifetime, and every later join token must match it or the request returns 409 — so set it here when you already know, to get that error on the room instead of on your first token. examples: - dual type: object title: CreateRoomRequest description: >- Request to create a translation room. Nothing is required. Languages are **not** set here — each participant declares their own when you request their join token, so a room can be created before anyone has chosen one. CreateRoomResponse: properties: id: type: string format: uuid title: Id room: type: string title: Room url: type: string title: Url streaming_mode: anyOf: - type: string - type: 'null' title: Streaming Mode type: object required: - id - room - url - streaming_mode title: CreateRoomResponse description: >- A created room, ready for participants. Creating a room does not issue any credentials — get a join token for each participant from `POST /v1/rooms/{room_name}/token`, which is also where that participant's languages are set. Use `id` to look the room and its usage up later. HTTPValidationError: properties: detail: items: $ref: '#/components/schemas/ValidationError' type: array title: Detail type: object title: HTTPValidationError StreamingMode: type: string enum: - mono - dual title: StreamingMode description: >- How a room's audio reaches the duplex model. A property of the **model's** wire contract, not a policy of ours — it supports these two modes and no others — which is why it lives here beside the rest of the API↔worker contract rather than in either deployable. The API validates a caller's choice against this enum at its boundary, exactly as it does for ``LanguageCode``; the worker will read the same values back off a dispatch, and one definition is what stops the two sides disagreeing about a mode's spelling. * ``MONO`` — both languages arrive on a single uplink, because both speakers share one device. * ``DUAL`` — each speaker has their own device, so each language arrives on its own track. ValidationError: properties: loc: items: anyOf: - type: string - type: integer type: array title: Location msg: type: string title: Message type: type: string title: Error Type input: title: Input ctx: type: object title: Context type: object required: - loc - msg - type title: ValidationError securitySchemes: HTTPBearer: type: http scheme: bearer ````