Together AI integration (AI, Developer Tools). Imported from open-connector.
https://api.together.ai/v1/models
/chat/completions
| Param | Type | Required | Description |
|---|---|---|---|
| model | string | required | The Together AI model identifier to use. |
| messages | array | required | The ordered conversation history sent to the model. |
| max_tokens | integer | optional | The maximum number of tokens to generate. |
| stop | array | optional | The stop sequences where generation should stop. |
| temperature | number | optional | The sampling temperature. |
| top_p | number | optional | The nucleus sampling threshold. |
| top_k | integer | optional | The maximum number of next-token choices to consider. |
| context_length_exceeded_behavior | string | optional | How the API should behave when max_tokens exceeds the model context length. |
| repetition_penalty | number | optional | The repetition penalty applied to generated text. |
| stream | boolean | optional | Whether to request a streaming response. This connector only accepts false or a… |
| logprobs | integer | optional | The number of top token log probabilities to include. |
| echo | boolean | optional | Whether the response should include the prompt. |
| n | integer | optional | The number of completions to generate for each prompt. |
| min_p | number | optional | An alternative probability threshold to top_p and top_k. |
| presence_penalty | number | optional | The presence penalty applied to newly introduced topics. |
| frequency_penalty | number | optional | The frequency penalty applied to repeated tokens. |
| logit_bias | object | optional | Token bias adjustments keyed by token id. |
| seed | integer | optional | A seed for reproducible sampling. |
| function_call | string | optional | How a legacy function call should be selected. |
| response_format | object | optional | Response format configuration. |
| tools | array | optional | Tools available to the model. |
| tool_choice | string | optional | Tool selection strategy for the request. |
| compliance | string | optional | The compliance mode accepted by the API. |
| chat_template_kwargs | object | optional | A JSON object with arbitrary upstream fields. |
| safety_model | string | optional | The moderation model used to validate tokens. |
| reasoning_effort | string | optional | The reasoning effort level to apply. |
| reasoning | object | optional | Reasoning configuration for models that support toggling reasoning. |
/embeddings
| Param | Type | Required | Description |
|---|---|---|---|
| model | string | optional | The Together AI embedding model identifier to use. |
| input | string | optional | The text input or inputs to embed. |