> ## Documentation Index > Fetch the complete documentation index at: https://docs.prem.io/llms.txt > Use this file to discover all available pages before exploring further. # Encrypted Chat Completions (RVENC) > Simple pass-through to enclave chat completion endpoint with encrypted payload. Rvenc = raw volatile encrypted. Accepts a single encrypted inference parameter that encapsulates all chat completion data. Only provides authentication, security checks, and token rate limiting. No chat history, file management, or custom features. # TypeScript SDK The SDK is available on GitHub: [premAI-io/api-sdk-ts](https://github.com/premAI-io/api-sdk-ts) ## Usage ### Basic Setup Create a client with auto-generated encryption keys: ```typescript theme={"system"} import createRvencClient from "@premai/api-sdk"; const client = await createRvencClient({ apiKey: process.env.API_KEY, clientKEK: process.env.CLIENT_KEK }); ``` ### Pre-generate Keys You can pre-generate encryption keys and reuse them: ```typescript theme={"system"} import createRvencClient, { generateEncryptionKeys } from "@premai/api-sdk"; const encryptionKeys = await generateEncryptionKeys(); const client = await createRvencClient({ apiKey: process.env.API_KEY, encryptionKeys, requestTimeoutMs: 60000, // optional maxBufferSize: 20 * 1024 * 1024, // optional }); ``` ### Non-streaming Requests ```typescript theme={"system"} const response = await client.chat.completions.create({ model: "openai/gpt-oss-120b", messages: [{ role: "user", content: "Hello!" }], }); ``` ### Streaming Requests ```typescript theme={"system"} const stream = await client.chat.completions.create({ model: "openai/gpt-oss-120b", messages: [{ role: "user", content: "Hello!" }], stream: true, }); for await (const chunk of stream) { process.stdout.write(chunk.choices[0]?.delta?.content || ""); } ``` ## Configuration | Option | Default | Description | | ------------------ | -------------- | -------------------------------------------- | | `apiKey` | required | Authorization token | | `encryptionKeys` | auto-generated | Pre-generated `{ sharedSecret, cipherText }` | | `requestTimeoutMs` | 30000 | Request timeout in ms | | `maxBufferSize` | 10MB | Max SSE buffer size | ## OpenAI-Compatible API Server Run as a standalone server with automatic DEK store management per API key: ```bash theme={"system"} npm start # Server runs on http://localhost:3000 ``` Use with any OpenAI-compatible client: ```bash theme={"system"} curl http://localhost:3000/v1/chat/completions \ -H "Authorization: Bearer your-api-key" \ -H "Content-Type: application/json" \ -d '{"model": "openai/gpt-oss-120b", "messages": [{"role": "user", "content": "Hello!"}], "stream": false}' ``` Or use the SDK directly in Node.js: ```typescript theme={"system"} import OpenAI from "openai"; const client = new OpenAI({ apiKey: process.env.API_KEY, baseURL: "http://localhost:3000/v1", }); const stream = await client.chat.completions.create({ model: "openai/gpt-oss-120b", messages: [{ role: "user", content: "Count to 10" }], stream: true, }); for await (const chunk of stream) { process.stdout.write(chunk.choices[0]?.delta?.content || ""); } ``` The server caches clients in memory per API key for better performance. *** # API Reference ## OpenAPI ````yaml post /rvenc/chat/completions openapi: 3.1.0 info: title: Prem API description: Full description of private and dev-availble API endpoints version: 1.0.0 termsOfService: https://prem.io/terms contact: name: API Support url: https://help.prem.io email: support@premai.io servers: - url: https://gateway.prem.io description: Production API server security: [] paths: /rvenc/chat/completions: post: tags: - Chats - dev-api summary: Encrypted Chat Completions (RVENC) description: >- Simple pass-through to enclave chat completion endpoint with encrypted payload. Rvenc = raw volatile encrypted. Accepts a single encrypted inference parameter that encapsulates all chat completion data. Only provides authentication, security checks, and token rate limiting. No chat history, file management, or custom features. operationId: createRvencChatCompletion requestBody: $ref: '#/components/requestBodies/createRvencChatCompletionRequest' responses: '200': $ref: '#/components/responses/createRvencChatCompletionResponse' '400': $ref: '#/components/responses/400' '401': $ref: '#/components/responses/401' '403': $ref: '#/components/responses/403' '429': description: Too many requests - concurrent stream limit reached content: application/json: schema: type: object properties: status: type: integer examples: - 429 error: type: string examples: - >- You already have an active chat stream. Please wait for it to complete or try again in a few moments. security: - BearerAuth: [] ApiKeyAuth: [] components: requestBodies: createRvencChatCompletionRequest: required: true content: application/json: schema: $ref: '#/components/schemas/createRvencChatCompletionRequest' responses: '400': description: Bad request content: application/json: schema: type: object properties: status: type: integer enum: - 400 - 401 - 403 - 404 description: Status code of the response data: type: - object - 'null' description: Response data containing the requested object error: type: - string - 'null' examples: - Some error message description: Error message of the response, human readable message: type: 'null' env: type: string enum: - development - production description: API environment log: type: - string - object - 'null' examples: - request_id: req_1234567890 - Some pertinent log message description: Useful informaiton, not always present, to debug the response validator: type: - object - array - 'null' examples: - email: Invalid email address password: Password is required description: >- Validator response object, each key is the field name and value is the error message support_id: type: - string - 'null' format: uuid examples: - support_uuidv7-something-else description: >- Support ID linked to the response, used to identify it when talking with our team required: - status - error - message - env - log - support_id additionalProperties: false '401': description: Access token is missing or invalid content: application/json: schema: allOf: - type: object properties: status: type: integer enum: - 400 - 401 - 403 - 404 description: Status code of the response data: type: - object - 'null' description: Response data containing the requested object error: type: - string - 'null' examples: - Some error message description: Error message of the response, human readable message: type: 'null' env: type: string enum: - development - production description: API environment log: type: - string - object - 'null' examples: - request_id: req_1234567890 - Some pertinent log message description: >- Useful informaiton, not always present, to debug the response validator: type: - object - array - 'null' examples: - email: Invalid email address password: Password is required description: >- Validator response object, each key is the field name and value is the error message support_id: type: - string - 'null' format: uuid examples: - support_uuidv7-something-else description: >- Support ID linked to the response, used to identify it when talking with our team required: - status - error - message - env - log - support_id additionalProperties: false - properties: status: type: integer enum: - 401 '403': description: You do not have the required permissions to access this resource content: application/json: schema: allOf: - type: object properties: status: type: integer enum: - 400 - 401 - 403 - 404 description: Status code of the response data: type: - object - 'null' description: Response data containing the requested object error: type: - string - 'null' examples: - Some error message description: Error message of the response, human readable message: type: 'null' env: type: string enum: - development - production description: API environment log: type: - string - object - 'null' examples: - request_id: req_1234567890 - Some pertinent log message description: >- Useful informaiton, not always present, to debug the response validator: type: - object - array - 'null' examples: - email: Invalid email address password: Password is required description: >- Validator response object, each key is the field name and value is the error message support_id: type: - string - 'null' format: uuid examples: - support_uuidv7-something-else description: >- Support ID linked to the response, used to identify it when talking with our team required: - status - error - message - env - log - support_id additionalProperties: false - properties: status: type: integer enum: - 403 createRvencChatCompletionResponse: description: >- Server-sent events stream with encrypted OpenAI-compatible chat completion chunks. Each chunk in the stream is encrypted and must be decrypted using the same shared secret and nonce from the request. content: text/event-stream: schema: $ref: '#/components/schemas/createRvencChatCompletionResponse' schemas: createRvencChatCompletionRequest: type: object description: >- Request body for rvenc (raw volatile encrypted) chat completion. Contains an encrypted payload with cryptographic materials needed for decryption. properties: encryptedInference: type: string description: >- Encrypted JSON string containing all chat completion parameters. When decrypted, this string must match the structure shown in the expandable `_decryptedInference` property below (reference only - do not send this property). _decryptedInference: $ref: '#/components/schemas/rvencInferenceParams' description: >- **Reference only - Do NOT send this property** - This shows the structure that `encryptedInference` should contain when decrypted. Encrypt a JSON object matching this structure and send it as the `encryptedInference` string. readOnly: true cipherText: type: string description: Cipher text for shared secret generation (ECDH key exchange) nonce: type: string description: Nonce used for encrypting the inference payload required: - encryptedInference - cipherText - nonce additionalProperties: false examples: - encryptedInference: encrypted_inference_payload_data cipherText: cipher_text_for_key_exchange nonce: nonce_value createRvencChatCompletionResponse: type: object description: >- Server-sent events (SSE) stream response structure. The actual response is a text stream, but this schema documents the structure for reference. properties: stream: type: string description: >- Server-sent events (SSE) stream with encrypted chat completion chunks. The stream contains three types of events: - `event: data` followed by `data: ` - Encrypted chunk that must be decrypted using the same shared secret and nonce from the request. When decrypted, each chunk matches the structure shown in the `_decryptedChunk` property below (reference only). - `event: error` followed by `data: ` - Error event (can be encrypted or plain JSON) - `event: done` followed by `data: [DONE]` - Stream completion marker Each encrypted `data:` line contains a hex-encoded string. Decrypt each chunk using XChaCha20-Poly1305 with the shared secret and nonce from your request. _decryptedChunk: $ref: '#/components/schemas/rvencChatCompletionChunk' description: >- **Reference only** - This shows the structure that each encrypted chunk should contain when decrypted. Decrypt each chunk in the stream using the same shared secret and nonce from your request to get this structure. readOnly: true required: - stream examples: - stream: |- event: data data: a1b2c3d4e5f67890abcdef1234567890abcdef123456... event: data data: f6e5d4c3b2a19876543210fedcba9876543210fedcba... event: data data: 1234567890abcdef1234567890abcdef1234567890... event: done data: [DONE] - stream: |- event: data data: a1b2c3d4e5f67890abcdef1234567890abcdef123456... event: error data: {"error":{"message":"Error occurred","type":"api_error"}} event: done data: [DONE] rvencInferenceParams: type: object description: >- Decrypted inference parameters that are encrypted and sent as encryptedInference. This schema documents the structure of the JSON that should be encrypted and included in the encryptedInference field. properties: messages: type: array description: >- A list of messages comprising the conversation so far. Depending on the model you use, different message types (modalities) are supported, like text, images, and audio. items: oneOf: - type: object description: >- Developer-provided instructions that the model should follow. With o1 models and newer, use developer messages for this purpose instead. properties: role: type: string enum: - system description: The role of the messages author, in this case system content: oneOf: - type: string description: The contents of the system message - type: array description: >- An array of content parts with a defined type. For system messages, only type text is supported items: type: object properties: text: type: string description: The text content type: type: string enum: - text description: The type of the content part required: - text - type description: The contents of the system message name: type: string description: >- An optional name for the participant. Provides the model information to differentiate between participants of the same role required: - role - content - type: object description: >- Messages sent by an end user, containing prompts or additional context information properties: role: type: string enum: - user description: The role of the messages author, in this case user content: oneOf: - type: string description: The text contents of the message - type: array description: >- An array of content parts with a defined type. Supported options differ based on the model being used to generate the response. Can contain text, image, or audio inputs items: oneOf: - type: object properties: text: type: string description: The text content type: type: string enum: - text description: The type of the content part required: - text - type - type: object properties: image_url: type: object properties: url: type: string description: >- Either a URL of the image or the base64 encoded image data detail: type: string enum: - low - high - auto default: auto description: Specifies the detail level of the image required: - url type: type: string enum: - image_url description: The type of the content part required: - image_url - type - type: object properties: input_audio: type: object properties: data: type: string description: Base64 encoded audio data format: type: string enum: - wav - mp3 description: The format of the encoded audio data required: - data - format type: type: string enum: - input_audio description: >- The type of the content part. Always input_audio required: - input_audio - type - type: object properties: file: type: object properties: file_data: type: string description: >- The base64 encoded file data, used when passing the file to the model as a string file_id: type: string description: >- The ID of an uploaded file to use as input filename: type: string description: >- The name of the file, used when passing the file to the model as a string type: type: string enum: - file description: The type of the content part. Always file required: - file - type description: The contents of the user message name: type: string description: >- An optional name for the participant. Provides the model information to differentiate between participants of the same role required: - role - content - type: object description: Messages sent by the model in response to user messages properties: role: type: string enum: - assistant description: The role of the messages author, in this case assistant audio: type: object properties: id: type: string description: >- Unique identifier for a previous audio response from the model required: - id description: Data about a previous audio response from the model content: oneOf: - type: string - type: array items: type: object properties: text: type: string description: The text content type: type: string enum: - text description: The type of the content part required: - text - type description: >- The contents of the assistant message. Required unless tool_calls or function_call is specified function_call: type: object properties: name: type: string description: The name of the function to call arguments: type: string description: >- The arguments to call the function with, as generated by the model in JSON format required: - name - arguments description: >- Deprecated and replaced by tool_calls. The name and arguments of a function that should be called, as generated by the model name: type: string description: >- An optional name for the participant. Provides the model information to differentiate between participants of the same role refusal: type: string description: The refusal message by the assistant tool_calls: type: array description: >- The tool calls generated by the model, such as function calls items: type: object properties: id: type: string description: Unique identifier for the tool call type: type: string description: Type of tool call function: type: object properties: name: type: string description: Name of the function to call arguments: type: string description: JSON string of function arguments required: - name - arguments required: - id - type - function required: - role - type: object description: Tool message properties: role: type: string enum: - tool description: The role of the messages author, in this case tool content: oneOf: - type: string description: The contents of the tool message - type: array description: >- An array of content parts with a defined type. For tool messages, only type text is supported items: type: object properties: text: type: string description: The text content type: type: string enum: - text description: The type of the content part required: - text - type description: The contents of the tool message tool_call_id: type: string description: Tool call that this message is responding to required: - role - content - tool_call_id - type: object description: Deprecated function message properties: role: type: string enum: - function description: The role of the messages author, in this case function content: type: string description: The contents of the function message name: type: string description: The name of the function to call required: - role - content - name minItems: 1 model: type: string description: The model identifier to use for completion temperature: type: number minimum: 0 maximum: 2 description: >- Sampling temperature between 0 and 2. Higher values make output more random. top_p: type: number minimum: 0 maximum: 1 description: >- Nucleus sampling parameter. Consider tokens with top_p probability mass. max_tokens: type: - integer - 'null' description: >- The maximum number of tokens that can be generated in the chat completion. This value can be used to control costs for text generated via API. This value is now deprecated in favor of max_completion_tokens, and is not compatible with o-series models max_completion_tokens: type: - integer - 'null' description: >- An upper bound for the number of tokens that can be generated for a completion, including visible output tokens and reasoning tokens stream: type: - boolean - 'null' default: false description: >- If set to true, the model response data will be streamed to the client as it is generated using server-sent events tools: type: array description: >- A list of tools the model may call. You can provide either custom tools or function tools items: oneOf: - type: object description: A function tool that can be used to generate a response properties: type: type: string enum: - function description: >- The type of the tool. Currently, only function is supported function: type: object properties: name: type: string description: >- The name of the function to be called. Must be a-z, A-Z, 0-9, or contain underscores and dashes, with a maximum length of 64 description: type: string description: >- A description of what the function does, used by the model to choose when and how to call the function parameters: type: object description: >- The parameters the functions accepts, described as a JSON Schema object. Omitting parameters defines a function with an empty parameter list required: - name required: - type - function - type: object description: A custom tool that processes input using a specified format properties: type: type: string enum: - custom description: The type of the custom tool. Always custom custom: type: object properties: name: type: string description: >- The name of the custom tool, used to identify it in tool calls description: type: string description: >- Optional description of the custom tool, used to provide more context format: type: object description: >- The input format for the custom tool. Default is unconstrained text required: - name description: Properties of the custom tool required: - type - custom tool_choice: oneOf: - type: string enum: - none - auto - required description: >- none means the model will not call any tool and instead generates a message. auto means the model can pick between generating a message or calling one or more tools. required means the model must call one or more tools - type: object properties: type: type: string enum: - function description: For function calling, the type is always function function: type: object properties: name: type: string description: The name of the function to call required: - name required: - type - function description: >- Specifies a tool the model should use. Use to force the model to call a specific function - type: object properties: type: type: string enum: - custom description: For custom tool calling, the type is always custom custom: type: object properties: name: type: string description: The name of the custom tool to call required: - name required: - type - custom description: >- Specifies a tool the model should use. Use to force the model to call a specific custom tool description: >- Controls which (if any) tool is called by the model. none means the model will not call any tool and instead generates a message. auto means the model can pick between generating a message or calling one or more tools. required means the model must call one or more tools. Specifying a particular tool forces the model to call that tool response_format: oneOf: - type: object properties: type: type: string enum: - text description: The type of response format being defined. Always text required: - type description: Default response format. Used to generate text responses - type: object properties: type: type: string enum: - json_schema description: >- The type of response format being defined. Always json_schema json_schema: type: object properties: name: type: string description: >- The name of the response format. Must be a-z, A-Z, 0-9, or contain underscores and dashes, with a maximum length of 64 description: type: string description: >- A description of what the response format is for, used by the model to determine how to respond in the format schema: type: object description: >- The schema for the response format, described as a JSON Schema object strict: type: boolean default: false description: >- Whether to enable strict schema adherence when generating the output. If set to true, the model will always follow the exact schema defined in the schema field required: - name description: >- Structured Outputs configuration options, including a JSON Schema required: - type - json_schema description: >- JSON Schema response format. Used to generate structured JSON responses - type: object properties: type: type: string enum: - json_object description: >- The type of response format being defined. Always json_object required: - type description: >- JSON object response format. An older method of generating JSON responses. Using json_schema is recommended for models that support it description: An object specifying the format that the model must output audio: type: - object - 'null' description: >- Parameters for audio output. Required when audio output is requested with modalities: ["audio"] properties: format: type: string enum: - wav - mp3 - flac - opus - pcm16 description: Specifies the output audio format voice: type: string enum: - alloy - ash - ballad - coral - echo - fable - nova - onyx - sage - shimmer description: The voice the model uses to respond required: - format - voice frequency_penalty: type: - number - 'null' default: 0 minimum: -2 maximum: 2 description: >- Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim function_call: oneOf: - type: string enum: - none - auto description: >- none means the model will not call a function and instead generates a message. auto means the model can pick between generating a message or calling a function - type: object properties: name: type: string description: The name of the function to call required: - name description: >- Specifying a particular function via {"name": "my_function"} forces the model to call that function description: >- Deprecated in favor of tool_choice. Controls which (if any) function is called by the model functions: type: array description: >- Deprecated in favor of tools. A list of functions the model may generate JSON inputs for items: type: object properties: name: type: string description: >- The name of the function to be called. Must be a-z, A-Z, 0-9, or contain underscores and dashes, with a maximum length of 64 description: type: string description: >- A description of what the function does, used by the model to choose when and how to call the function parameters: type: object description: >- The parameters the functions accepts, described as a JSON Schema object. Omitting parameters defines a function with an empty parameter list required: - name logit_bias: type: object additionalProperties: type: number minimum: -100 maximum: 100 description: >- Modify the likelihood of specified tokens appearing in the completion. Accepts a JSON object that maps tokens (specified by their token ID in the tokenizer) to an associated bias value from -100 to 100 logprobs: type: - boolean - 'null' default: false description: >- Whether to return log probabilities of the output tokens or not. If true, returns the log probabilities of each output token returned in the content of message metadata: type: object additionalProperties: type: string maxLength: 512 maxProperties: 16 description: >- Set of 16 key-value pairs that can be attached to an object. Keys are strings with a maximum length of 64 characters. Values are strings with a maximum length of 512 characters modalities: type: array description: Output types that you would like the model to generate items: type: string enum: - text - audio default: - text 'n': type: - integer - 'null' default: 1 minimum: 1 description: >- How many chat completion choices to generate for each input message. Note that you will be charged based on the number of generated tokens across all of the choices. Keep n as 1 to minimize costs prediction: type: object description: >- Configuration for a Predicted Output, which can greatly improve response times when large parts of the model response are known ahead of time properties: type: type: string enum: - content description: >- The type of the predicted content you want to provide. This type is currently always content content: oneOf: - type: string description: >- The content used for a Predicted Output. This is often the text of a file you are regenerating with minor changes - type: array description: >- An array of content parts with a defined type. Supported options differ based on the model being used to generate the response. Can contain text inputs items: type: object properties: text: type: string description: The text content type: type: string enum: - text description: The type of the content part required: - text - type description: >- The content that should be matched when generating a model response required: - type - content presence_penalty: type: - number - 'null' default: 0 minimum: -2 maximum: 2 description: >- Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics prompt_cache_key: type: string description: >- Used by OpenAI to cache responses for similar requests to optimize your cache hit rates. Replaces the user field prompt_cache_retention: type: string description: >- The retention policy for the prompt cache. Set to 24h to enable extended prompt caching, which keeps cached prefixes active for longer, up to a maximum of 24 hours reasoning_effort: type: string enum: - none - minimal - low - medium - high default: medium description: >- Constrains effort on reasoning for reasoning models. Currently supported values are none, minimal, low, medium, and high. Reducing reasoning effort can result in faster responses and fewer tokens used on reasoning in a response safety_identifier: type: string description: >- A stable identifier used to help detect users of your application that may be violating OpenAI's usage policies. The IDs should be a string that uniquely identifies each user seed: type: - integer - 'null' description: >- This feature is in Beta. If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed service_tier: type: string enum: - auto - default - flex - priority default: auto description: Specifies the processing type used for serving the request stop: oneOf: - type: string - type: array items: type: string maxItems: 4 - type: 'null' default: null description: >- Up to 4 sequences where the API will stop generating further tokens. The returned text will not contain the stop sequence. Not supported with latest reasoning models o3 and o4-mini store: type: - boolean - 'null' default: false description: >- Whether or not to store the output of this chat completion request for use in our model distillation or evals products. Supports text and image inputs. Note: image inputs over 8MB will be dropped stream_options: type: object description: >- Options for streaming response. Only set this when you set stream: true properties: include_obfuscation: type: boolean description: >- When true, stream obfuscation will be enabled. Stream obfuscation adds random characters to an obfuscation field on streaming delta events to normalize payload sizes as a mitigation to certain side-channel attacks include_usage: type: boolean description: >- If set, an additional chunk will be streamed before the data: [DONE] message. The usage field on this chunk shows the token usage statistics for the entire request top_logprobs: type: integer minimum: 0 maximum: 20 description: >- An integer between 0 and 20 specifying the number of most likely tokens to return at each token position, each with an associated log probability verbosity: type: string enum: - low - medium - high default: medium description: >- Constrains the verbosity of the model's response. Lower values will result in more concise responses, while higher values will result in more verbose responses web_search_options: type: object description: This tool searches the web for relevant results to use in a response properties: search_context_size: type: string enum: - low - medium - high default: medium description: >- High level guidance for the amount of context window space to use for the search user_location: oneOf: - type: object properties: approximate: type: object properties: city: type: string description: >- Free text input for the city of the user, e.g. San Francisco country: type: string description: The two-letter ISO country code of the user, e.g. US region: type: string description: >- Free text input for the region of the user, e.g. California timezone: type: string description: >- The IANA timezone of the user, e.g. America/Los_Angeles type: type: string enum: - approximate description: The type of location approximation. Always approximate required: - approximate - type - type: 'null' description: Approximate location parameters for the search required: - messages - model additionalProperties: false examples: - messages: - role: user content: Hello, how are you? model: openai/gpt-oss-120b temperature: 0.7 stream: true - messages: - role: system content: You are a helpful assistant. - role: user content: What's the weather like? model: openai/gpt-oss-120b temperature: 0.8 top_p: 0.9 max_tokens: 500 stream: true tools: - type: function function: name: get_weather description: Get the current weather for a location parameters: type: object properties: location: type: string description: The city and state, e.g. San Francisco, CA required: - location tool_choice: auto rvencChatCompletionChunk: type: object description: >- Decrypted chat completion chunk structure that matches OpenAI's streaming format. Each encrypted chunk in the SSE stream, when decrypted, matches this structure. properties: id: type: string description: >- A unique identifier for the chat completion. Each chunk has the same id. examples: - cd9d05b657a041a6a14ab2fc890a7d7e object: type: string enum: - chat.completion.chunk description: The object type, which is always 'chat.completion.chunk' examples: - chat.completion.chunk created: type: integer description: >- The Unix timestamp (in seconds) of when the chat completion was created examples: - 1764004858 model: type: string description: The model used for the chat completion examples: - openai/gpt-oss-120b system_fingerprint: type: - string - 'null' description: >- This fingerprint represents the backend configuration that the model runs with (optional) examples: - fp_44709d6fcb choices: type: array description: >- A list of chat completion choices. Can contain more than one element if n is greater than 1. items: type: object properties: index: type: integer description: The index of the choice in the list of choices examples: - 0 delta: type: object description: >- A chat completion delta generated by streamed model responses. The delta can contain 'role', 'content', 'reasoning_content', 'tool_calls', or be empty {} when finish_reason is set. properties: role: type: - string - 'null' enum: - assistant - null description: >- The role of the author of this message (only present in first chunk, can be null) examples: - assistant content: type: - string - 'null' description: >- The contents of the chunk message (can be null or empty string) examples: - Hello reasoning_content: type: - string - 'null' description: >- Reasoning content for models that support reasoning (optional, can be null) examples: - ':' tool_calls: type: - array - 'null' description: Tool calls made by the assistant (optional, can be null) items: type: object logprobs: type: - object - 'null' description: Log probability information for the choice default: null finish_reason: type: - string - 'null' enum: - stop - length - tool_calls - content_filter - null description: >- The reason the model stopped generating tokens. This will be null if the model hasn't finished generating the tokens yet. examples: - stop matched_stop: type: - string - 'null' description: The stop sequence that was matched (optional, can be null) default: null required: - index - delta - logprobs - finish_reason usage: type: - object - 'null' description: Usage statistics for the completion (optional, can be null) properties: prompt_tokens: type: integer description: Number of tokens in the prompt completion_tokens: type: integer description: Number of tokens in the completion total_tokens: type: integer description: Total number of tokens used default: null error: type: object description: Error object if an error occurred (optional) properties: message: type: string description: Error message type: type: string description: Error type required: - id - object - created - model - choices additionalProperties: false examples: - id: cd9d05b657a041a6a14ab2fc890a7d7e object: chat.completion.chunk created: 1764004858 model: openai/gpt-oss-120b choices: - index: 0 delta: role: null content: null reasoning_content: 'This ' tool_calls: null logprobs: null finish_reason: null matched_stop: null usage: null - id: cd9d05b657a041a6a14ab2fc890a7d7e object: chat.completion.chunk created: 1764004858 model: openai/gpt-oss-120b choices: - index: 0 delta: role: assistant content: Hello reasoning_content: null tool_calls: null logprobs: null finish_reason: null matched_stop: null usage: null - id: cd9d05b657a041a6a14ab2fc890a7d7e object: chat.completion.chunk created: 1764004858 model: openai/gpt-oss-120b choices: - index: 0 delta: [] logprobs: null finish_reason: stop matched_stop: null usage: null securitySchemes: BearerAuth: type: http scheme: bearer bearerFormat: JWT description: 'Send your access token as header Authorization: Bearer {accessToken}' ApiKeyAuth: type: apiKey in: header name: Authorization description: >- Your API key that starts with sk_live or sk_test. You can create yours at go.prem.io/api-keys. ````