--- title: "Create a new Embedding Model Configuration" method: POST path: "/api/v1/embedding-model-configs" tags: ["Embedding Model Configs"] --- # Create a new Embedding Model Configuration `POST /api/v1/embedding-model-configs` Create a new embedding model configuration within a specified project. ## Query parameters - `project_id` string, uuid, nullable - `organization_id` string, uuid, nullable ## Cookies - `session` string, nullable ## Request body - EmbeddingModelConfigCreate - `name` string, required — The name of the embedding model config. - `embedding_config` union, required — The embedding configuration for the embedding model config. - AzureOpenAIEmbeddingConfig - `type` 'AZURE_EMBEDDING' — Type of the embedding model. - `component` AzureOpenAIEmbedding - `model_name` string — The name of the OpenAI embedding model. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `additional_kwargs` object — Additional kwargs for the OpenAI API. - `api_key` string, nullable — The OpenAI API key. - `api_base` string — The base URL for Azure deployment. - `api_version` string — The version for Azure OpenAI API. - `max_retries` integer — Maximum number of retries. - `timeout` number — Timeout for each request. - `default_headers` object, nullable — The default headers for API requests. - `reuse_client` boolean — Reuse the OpenAI client between requests. When doing anything with large volumes of async API calls, setting this to false can improve stability. - `dimensions` integer, nullable — The number of dimensions on the output embedding vectors. Works only with v3 embedding models. - `azure_endpoint` string, nullable — The Azure endpoint to use. - `azure_deployment` string, nullable — The Azure deployment to use. - `class_name` string - CohereEmbeddingConfig - `type` 'COHERE_EMBEDDING' — Type of the embedding model. - `component` CohereEmbedding - `model_name` string — The modelId of the Cohere model to use. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `api_key` string, nullable, required — The Cohere API key. - `truncate` string — Truncation type - START/ END/ NONE - `input_type` string, nullable — Model Input type. If not provided, search_document and search_query are used when needed. - `embedding_type` string — Embedding type. If not provided float embedding_type is used when needed. - `class_name` string - GeminiEmbeddingConfig - `type` 'GEMINI_EMBEDDING' — Type of the embedding model. - `component` GeminiEmbedding - `model_name` string — The modelId of the Gemini model to use. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `title` string, nullable — Title is only applicable for retrieval_document tasks, and is used to represent a document title. For other tasks, title is invalid. - `task_type` string, nullable — The task for embedding model. - `api_key` string, nullable — API key to access the model. Defaults to None. - `api_base` string, nullable — API base to access the model. Defaults to None. - `transport` string, nullable — Transport to access the model. Defaults to None. - `output_dimensionality` integer, nullable — Optional reduced dimension for output embeddings. Supported by models/text-embedding-004 and newer (e.g. gemini-embedding-001). Not supported by models/embedding-001. - `class_name` string - HuggingFaceInferenceAPIEmbeddingConfig - `type` 'HUGGINGFACE_API_EMBEDDING' — Type of the embedding model. - `component` HuggingFaceInferenceAPIEmbedding - `model_name` string, nullable — Hugging Face model name. If None, the task will be used. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `pooling` 'cls' | 'mean' | 'last' — Enum of possible pooling choices with pooling behaviors. - `query_instruction` string, nullable — Instruction to prepend during query embedding. - `text_instruction` string, nullable — Instruction to prepend during text embedding. - `token` union — Hugging Face token. Will default to the locally saved token. Pass token=False if you don’t want to send your token to the server. - string - boolean - `timeout` number, nullable — The maximum number of seconds to wait for a response from the server. Loading a new model in Inference API can take up to several minutes. Defaults to None, meaning it will loop until the server is available. - `headers` object, nullable — Additional headers to send to the server. By default only the authorization and user-agent headers are sent. Values in this dictionary will override the default values. - `cookies` object, nullable — Additional cookies to send to the server. - `task` string, nullable — Optional task to pick Hugging Face's recommended model, used when model_name is left as default of None. - `class_name` string - OpenAIEmbeddingConfig - `type` 'OPENAI_EMBEDDING' — Type of the embedding model. - `component` OpenAIEmbedding - `model_name` string — The name of the OpenAI embedding model. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `additional_kwargs` object — Additional kwargs for the OpenAI API. - `api_key` string, nullable — The OpenAI API key. - `api_base` string, nullable — The base URL for OpenAI API. - `api_version` string, nullable — The version for OpenAI API. - `max_retries` integer — Maximum number of retries. - `timeout` number — Timeout for each request. - `default_headers` object, nullable — The default headers for API requests. - `reuse_client` boolean — Reuse the OpenAI client between requests. When doing anything with large volumes of async API calls, setting this to false can improve stability. - `dimensions` integer, nullable — The number of dimensions on the output embedding vectors. Works only with v3 embedding models. - `class_name` string - VertexAIEmbeddingConfig - `type` 'VERTEXAI_EMBEDDING' — Type of the embedding model. - `component` VertexTextEmbedding - `model_name` string — The modelId of the VertexAI model to use. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `location` string, required — The default location to use when making API calls. - `project` string, required — The default GCP project to use when making Vertex API calls. - `embed_mode` 'default' | 'classification' | 'clustering' | 'similarity' | 'retrieval' — Copied from llama_index.embeddings.vertex.base.VertexEmbeddingMode since importing llama_index.embeddings.vertex.base incurs a lot of memory usage. - `additional_kwargs` object — Additional kwargs for the Vertex. - `client_email` string, nullable, required — The client email for the VertexAI credentials. - `token_uri` string, nullable, required — The token URI for the VertexAI credentials. - `private_key_id` string, nullable, required — The private key ID for the VertexAI credentials. - `private_key` string, nullable, required — The private key for the VertexAI credentials. - `class_name` string - BedrockEmbeddingConfig - `type` 'BEDROCK_EMBEDDING' — Type of the embedding model. - `component` BedrockEmbedding - `model_name` string — The modelId of the Bedrock model to use. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `profile_name` string, nullable — The name of aws profile to use. If not given, then the default profile is used. - `aws_access_key_id` string, nullable — AWS Access Key ID to use - `aws_secret_access_key` string, nullable — AWS Secret Access Key to use - `aws_session_token` string, nullable — AWS Session Token to use - `region_name` string, nullable — AWS region name to use. Uses region configured in AWS CLI if not passed - `max_retries` integer — The maximum number of API retries. - `timeout` number — The timeout for the Bedrock API request in seconds. It will be used for both connect and read timeouts. - `additional_kwargs` object — Additional kwargs for the bedrock client. - `class_name` string ## Response `201` Successful Response - EmbeddingModelConfig — Schema for an embedding model config. - `id` string, uuid, required — Unique identifier - `created_at` string, date-time, nullable — Creation datetime - `updated_at` string, date-time, nullable — Update datetime - `name` string, required — The name of the embedding model config. - `embedding_config` union, required — The embedding configuration for the embedding model config. - AzureOpenAIEmbeddingConfig - `type` 'AZURE_EMBEDDING' — Type of the embedding model. - `component` AzureOpenAIEmbedding - `model_name` string — The name of the OpenAI embedding model. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `additional_kwargs` object — Additional kwargs for the OpenAI API. - `api_key` string, nullable — The OpenAI API key. - `api_base` string — The base URL for Azure deployment. - `api_version` string — The version for Azure OpenAI API. - `max_retries` integer — Maximum number of retries. - `timeout` number — Timeout for each request. - `default_headers` object, nullable — The default headers for API requests. - `reuse_client` boolean — Reuse the OpenAI client between requests. When doing anything with large volumes of async API calls, setting this to false can improve stability. - `dimensions` integer, nullable — The number of dimensions on the output embedding vectors. Works only with v3 embedding models. - `azure_endpoint` string, nullable — The Azure endpoint to use. - `azure_deployment` string, nullable — The Azure deployment to use. - `class_name` string - CohereEmbeddingConfig - `type` 'COHERE_EMBEDDING' — Type of the embedding model. - `component` CohereEmbedding - `model_name` string — The modelId of the Cohere model to use. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `api_key` string, nullable, required — The Cohere API key. - `truncate` string — Truncation type - START/ END/ NONE - `input_type` string, nullable — Model Input type. If not provided, search_document and search_query are used when needed. - `embedding_type` string — Embedding type. If not provided float embedding_type is used when needed. - `class_name` string - GeminiEmbeddingConfig - `type` 'GEMINI_EMBEDDING' — Type of the embedding model. - `component` GeminiEmbedding - `model_name` string — The modelId of the Gemini model to use. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `title` string, nullable — Title is only applicable for retrieval_document tasks, and is used to represent a document title. For other tasks, title is invalid. - `task_type` string, nullable — The task for embedding model. - `api_key` string, nullable — API key to access the model. Defaults to None. - `api_base` string, nullable — API base to access the model. Defaults to None. - `transport` string, nullable — Transport to access the model. Defaults to None. - `output_dimensionality` integer, nullable — Optional reduced dimension for output embeddings. Supported by models/text-embedding-004 and newer (e.g. gemini-embedding-001). Not supported by models/embedding-001. - `class_name` string - HuggingFaceInferenceAPIEmbeddingConfig - `type` 'HUGGINGFACE_API_EMBEDDING' — Type of the embedding model. - `component` HuggingFaceInferenceAPIEmbedding - `model_name` string, nullable — Hugging Face model name. If None, the task will be used. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `pooling` 'cls' | 'mean' | 'last' — Enum of possible pooling choices with pooling behaviors. - `query_instruction` string, nullable — Instruction to prepend during query embedding. - `text_instruction` string, nullable — Instruction to prepend during text embedding. - `token` union — Hugging Face token. Will default to the locally saved token. Pass token=False if you don’t want to send your token to the server. - string - boolean - `timeout` number, nullable — The maximum number of seconds to wait for a response from the server. Loading a new model in Inference API can take up to several minutes. Defaults to None, meaning it will loop until the server is available. - `headers` object, nullable — Additional headers to send to the server. By default only the authorization and user-agent headers are sent. Values in this dictionary will override the default values. - `cookies` object, nullable — Additional cookies to send to the server. - `task` string, nullable — Optional task to pick Hugging Face's recommended model, used when model_name is left as default of None. - `class_name` string - OpenAIEmbeddingConfig - `type` 'OPENAI_EMBEDDING' — Type of the embedding model. - `component` OpenAIEmbedding - `model_name` string — The name of the OpenAI embedding model. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `additional_kwargs` object — Additional kwargs for the OpenAI API. - `api_key` string, nullable — The OpenAI API key. - `api_base` string, nullable — The base URL for OpenAI API. - `api_version` string, nullable — The version for OpenAI API. - `max_retries` integer — Maximum number of retries. - `timeout` number — Timeout for each request. - `default_headers` object, nullable — The default headers for API requests. - `reuse_client` boolean — Reuse the OpenAI client between requests. When doing anything with large volumes of async API calls, setting this to false can improve stability. - `dimensions` integer, nullable — The number of dimensions on the output embedding vectors. Works only with v3 embedding models. - `class_name` string - VertexAIEmbeddingConfig - `type` 'VERTEXAI_EMBEDDING' — Type of the embedding model. - `component` VertexTextEmbedding - `model_name` string — The modelId of the VertexAI model to use. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `location` string, required — The default location to use when making API calls. - `project` string, required — The default GCP project to use when making Vertex API calls. - `embed_mode` 'default' | 'classification' | 'clustering' | 'similarity' | 'retrieval' — Copied from llama_index.embeddings.vertex.base.VertexEmbeddingMode since importing llama_index.embeddings.vertex.base incurs a lot of memory usage. - `additional_kwargs` object — Additional kwargs for the Vertex. - `client_email` string, nullable, required — The client email for the VertexAI credentials. - `token_uri` string, nullable, required — The token URI for the VertexAI credentials. - `private_key_id` string, nullable, required — The private key ID for the VertexAI credentials. - `private_key` string, nullable, required — The private key for the VertexAI credentials. - `class_name` string - BedrockEmbeddingConfig - `type` 'BEDROCK_EMBEDDING' — Type of the embedding model. - `component` BedrockEmbedding - `model_name` string — The modelId of the Bedrock model to use. - `embed_batch_size` integer — The batch size for embedding calls. - `num_workers` integer, nullable — The number of workers to use for async embedding calls. - `profile_name` string, nullable — The name of aws profile to use. If not given, then the default profile is used. - `aws_access_key_id` string, nullable — AWS Access Key ID to use - `aws_secret_access_key` string, nullable — AWS Secret Access Key to use - `aws_session_token` string, nullable — AWS Session Token to use - `region_name` string, nullable — AWS region name to use. Uses region configured in AWS CLI if not passed - `max_retries` integer — The maximum number of API retries. - `timeout` number — The timeout for the Bedrock API request in seconds. It will be used for both connect and read timeouts. - `additional_kwargs` object — Additional kwargs for the bedrock client. - `class_name` string - `project_id` string, uuid, required ## Other responses - `422` — Validation Error --- [API](https://skmtc.net/run-llama/apis/llama-platform.md) · [All operations](https://skmtc.net/run-llama/apis/llama-platform/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/run-llama/llama-platform/revisions/b17341164de9/schema)