> ## Documentation Index > Fetch the complete documentation index at: https://docs.gbase.ai/llms.txt > Use this file to discover all available pages before exploring further. # Save Realtime Session > Batch save realtime voice session from live-stream WebSocket. Authentication: MD5 hash of "master:" + botId This endpoint accepts a batch of messages from a realtime voice conversation and saves them to the database with proper session and message tracking. ## OpenAPI ````yaml https://api.gbase.ai/openapi-redoc.json post /v1/robots/{ai_id}/realtime-session/save openapi: 3.1.0 info: title: GBase description: > # Introduce This API document can help you understand GBase more deeply. You can train your own AI directly through the interface of this document. ## Authentication Home page: [https://gbase.ai](https://gbase.ai) Huggingface: [https://huggingface.co/gptbase](https://huggingface.co/gptbase) The GBase API uses API keys for authentication. Visit your API Keys page to retrieve the API key you'll use in your requests. Remember that your API key is a secret! Do not share it with others or expose it in any client-side code (browsers, apps). Production requests must be routed through your own backend server where your API key can be securely loaded from an environment variable or key management service. All API requests should include your API key in an Authorization HTTP header as follows: Parameter name: Authorization Value: Bearer [YOUR_API_KEY](https://admin.gbase.ai/keys) ``` Authorization: Bearer YOUR_API_KEY ``` ## Use steps 1.Fill in your API keys. 2.You can create an AI by using either the API interface or the administrative backend. Once created, the AI will generate an ID which will be required for the subsequent steps. 3.Use the generated ID to call the file upload interface. 4.After uploading the file, you can use the question interface to inquire about it. version: 0.0.1 servers: - url: https://api.gbase.ai security: [] paths: /v1/robots/{ai_id}/realtime-session/save: post: tags: - Question summary: Save Realtime Session description: >- Batch save realtime voice session from live-stream WebSocket. Authentication: MD5 hash of "master:" + botId This endpoint accepts a batch of messages from a realtime voice conversation and saves them to the database with proper session and message tracking. operationId: save_realtime_session_v1_robots__ai_id__realtime_session_save_post parameters: - required: true schema: type: string title: Ai Id name: ai_id in: path requestBody: content: application/json: schema: $ref: '#/components/schemas/RealtimeSessionSaveIn' required: true responses: '200': description: Successful Response content: application/json: schema: $ref: '#/components/schemas/RealtimeSessionSaveOut' '422': description: Validation Error content: application/json: schema: $ref: '#/components/schemas/HTTPValidationError' security: - HTTPBearer: [] components: schemas: RealtimeSessionSaveIn: properties: session_id: type: string maxLength: 100 title: Session Id messages: items: $ref: '#/components/schemas/RealtimeMessageIn' type: array title: Messages metadata: type: object title: Metadata default: {} usage: $ref: '#/components/schemas/RealtimeUsageIn' tool_usage: $ref: '#/components/schemas/RealtimeToolUsageIn' logs: items: $ref: '#/components/schemas/RealtimeLogEntryIn' type: array title: Logs type: object required: - session_id - messages title: RealtimeSessionSaveIn RealtimeSessionSaveOut: properties: success: type: boolean title: Success session_id: type: string title: Session Id messages_saved: type: integer title: Messages Saved backend_session_id: type: string title: Backend Session Id type: object required: - success - session_id - messages_saved - backend_session_id title: RealtimeSessionSaveOut HTTPValidationError: properties: detail: items: $ref: '#/components/schemas/ValidationError' type: array title: Detail type: object title: HTTPValidationError RealtimeMessageIn: properties: role: type: string enum: - user - assistant title: Role content: type: string title: Content timestamp: type: string title: Timestamp message_type: type: string enum: - audio_transcription - text title: Message Type tool_calls: items: type: object type: array title: Tool Calls type: object required: - role - content - timestamp - message_type title: RealtimeMessageIn RealtimeUsageIn: properties: turn_count: type: integer title: Turn Count default: 0 total_tokens: type: integer title: Total Tokens default: 0 input_tokens: type: integer title: Input Tokens default: 0 output_tokens: type: integer title: Output Tokens default: 0 input_token_details: $ref: '#/components/schemas/RealtimeTokenDetailsIn' cached_tokens_details: $ref: '#/components/schemas/RealtimeTokenDetailsIn' output_token_details: $ref: '#/components/schemas/RealtimeTokenDetailsIn' turns: items: $ref: '#/components/schemas/RealtimeTurnUsageIn' type: array title: Turns default: [] type: object title: RealtimeUsageIn RealtimeToolUsageIn: properties: call_count: type: integer title: Call Count default: 0 total_prompt_tokens: type: integer title: Total Prompt Tokens default: 0 total_completion_tokens: type: integer title: Total Completion Tokens default: 0 total_tokens: type: integer title: Total Tokens default: 0 calls: items: $ref: '#/components/schemas/RealtimeToolCallUsageIn' type: array title: Calls default: [] type: object title: RealtimeToolUsageIn description: Realtime 会话级 tool LLM 用量聚合 RealtimeLogEntryIn: properties: timestamp: type: string title: Timestamp default: '' level: type: string title: Level default: log category: type: string title: Category default: '' message: type: string title: Message default: '' data: title: Data type: object title: RealtimeLogEntryIn description: 前端上报的 Realtime 日志条目 ValidationError: properties: loc: items: anyOf: - type: string - type: integer type: array title: Location msg: type: string title: Message type: type: string title: Error Type type: object required: - loc - msg - type title: ValidationError RealtimeTokenDetailsIn: properties: text_tokens: type: integer title: Text Tokens default: 0 audio_tokens: type: integer title: Audio Tokens default: 0 cached_tokens: type: integer title: Cached Tokens default: 0 type: object title: RealtimeTokenDetailsIn RealtimeTurnUsageIn: properties: turn_index: type: integer title: Turn Index total_tokens: type: integer title: Total Tokens default: 0 input_tokens: type: integer title: Input Tokens default: 0 output_tokens: type: integer title: Output Tokens default: 0 input_text_tokens: type: integer title: Input Text Tokens default: 0 input_audio_tokens: type: integer title: Input Audio Tokens default: 0 cached_tokens: type: integer title: Cached Tokens default: 0 cached_text_tokens: type: integer title: Cached Text Tokens default: 0 cached_audio_tokens: type: integer title: Cached Audio Tokens default: 0 output_text_tokens: type: integer title: Output Text Tokens default: 0 output_audio_tokens: type: integer title: Output Audio Tokens default: 0 duration_ms: type: integer title: Duration Ms default: 0 timestamp: type: integer title: Timestamp default: 0 latency_ms: type: integer title: Latency Ms latency_tool_call_count: type: integer title: Latency Tool Call Count default: 0 type: object required: - turn_index title: RealtimeTurnUsageIn RealtimeToolCallUsageIn: properties: turn_index: type: integer title: Turn Index default: 0 tool_name: type: string title: Tool Name default: '' model: type: string title: Model prompt_tokens: type: integer title: Prompt Tokens default: 0 completion_tokens: type: integer title: Completion Tokens default: 0 total_tokens: type: integer title: Total Tokens default: 0 timestamp: type: integer title: Timestamp default: 0 type: object title: RealtimeToolCallUsageIn description: Realtime 会话中单次工具调用产生的文本 LLM 用量（reasoning agent 等） securitySchemes: HTTPBearer: type: http scheme: bearer ````