Create an JinaAI inference endpoint
- PUT /_inference/{task_type}/{jinaai_inference_id}
Create an inference endpoint to perform an inference task with the jinaai service.
To review the available rerank models, refer to https://jina.ai/reranker.
To review the available embedding and text_embedding models, refer to https://jina.ai/embeddings/.
- Cluster privileges:
manage_inference
Path Parameters
-
task_typeJinaAITaskTyperequired -
The type of the inference task that the model will perform.
-
Values:
embeddingreranktext_embedding -
jinaai_inference_idstring Idrequired -
The unique identifier of the inference endpoint.
Query Parameters
-
timeoutstring Duration · default: 30s Specifies the amount of time to wait for the inference endpoint to be created.
Request
-
chunking_settingsInferenceChunkingSettings -
The chunking configuration object. Applies only to the
embeddingand text_embeddingtask types. Not applicable to thererank` task type. - Reference
-
max_chunk_sizenumber · default: 250 -
The maximum size of a chunk in words. This value cannot be lower than
20(forsentencestrategy) or10(forwordstrategy). This value should not exceed the window size for the associated model. -
overlapnumber · default: 100 -
The number of overlapping words for chunks. It is applicable only to a
wordchunking strategy. This value cannot be higher than half themax_chunk_sizevalue. -
sentence_overlapnumber · default: 1 -
The number of overlapping sentences for chunks. It is applicable only for a
sentencechunking strategy. It can be either1or0. -
separator_groupstring -
Only applicable to the
recursivestrategy and required when using it.Sets a predefined list of separators in the saved chunking settings based on the selected text type. Values can be
markdownorplaintext.Using this parameter is an alternative to manually specifying a custom
separatorslist. -
separatorsstring -
Only applicable to the
recursivestrategy and required when using it.A list of strings used as possible split points when chunking text.
Each string can be a plain string or a regular expression (regex) pattern. The system tries each separator in order to split the text, starting from the first item in the list.
After splitting, it attempts to recombine smaller pieces into larger chunks that stay within the
max_chunk_sizelimit, to reduce the total number of chunks generated. -
serviceJinaAIServiceTyperequired -
The type of service supported for the specified task type. In this case,
jinaai. -
Values:
jinaai -
service_settingsJinaAIServiceSettingsrequired -
Settings used to install the inference model. These settings are specific to the
jinaaiservice. -
api_keystringrequired -
A valid API key of your JinaAI account.
IMPORTANT: You need to provide the API key only once, during the inference model creation. The get inference endpoint API does not retrieve your API key.
- External docs
-
model_idstringrequired -
The name of the model to use for the inference task.
-
rate_limitRateLimitSetting -
This setting helps to minimize the number of rate limit errors returned from JinaAI. By default, the
jinaaiservice sets the number of requests allowed per minute to 2000 for all task types. - External docs
-
requests_per_minutenumber -
The number of requests allowed per minute. By default, the number of requests allowed per minute is set by each service as follows:
alibabacloud-ai-searchservice:1000amazonbedrockservice:240anthropicservice:50azureaistudioservice:240azureopenaiservice and task typetext_embedding:1440azureopenaiservice and task typescompletionorchat_completion:120cohereservice:10000contextualaiservice:1000elasticservice and task typechat_completion:240fireworksaiservice:6000googleaistudioservice:360googlevertexaiservice:30000hugging_faceservice:3000jinaaiservice:2000llamaservice:3000mistralservice:240openaiservice and task typetext_embedding:3000openaiservice and task typecompletion:500openshift_aiservice:3000voyageaiservice:2000watsonxaiservice:120
-
similarityJinaAISimilarityType -
For an
embeddingortext_embeddingtask, the similarity measure. One of cosine, dot_product, l2_norm. The default values varies with the embedding type. For example, a float embedding type uses adot_productsimilarity measure by default. -
Values:
cosinedot_productl2_norm -
dimensionsnumber -
For an
embeddingortext_embeddingtask, the number of dimensions the resulting output embeddings should have. By default, the model's standard output dimension is used. Refer to the Jina documentation for more information. - External docs
-
embedding_typeJinaAIElementType · default: float -
For an
embeddingortext_embeddingtask, the data type returned by the model. Usebitfor binary embeddings, which are encoded as bytes with signed int8 precision. Usebinaryfor binary embeddings, which are encoded as bytes with signed int8 precision (this is a synonym ofbit). Usefloatfor the default float embeddings. -
Values:
binarybitfloat -
multimodal_modelboolean · default: true -
For the
embeddingtask, whether the model supports multimodal inputs. If true, requests sent to the Jina model will use the multimodal request format (a list of objects). If false, requests sent to the model will use the same format as thetext_embeddingtask (a list of strings). Setting this tofalseallows theembeddingtask to be used with models that do not support multimodal requests. -
task_settingsJinaAITaskSettings -
Settings to configure the inference task. These settings are specific to the task type you specified.
-
return_documentsboolean -
For a
reranktask, return the doc text within the results. -
input_typeJinaAITextEmbeddingTask -
For an
embeddingortext_embeddingtask, the task passed to the model. Valid values are:classification: Use it for embeddings passed through a classifier.clustering: Use it for the embeddings run through a clustering algorithm.ingest: Use it for storing document embeddings in a vector database.search: Use it for storing embeddings of search queries run against a vector database to find relevant documents.
-
Values:
classificationclusteringingestsearch -
late_chunkingboolean -
For an
embeddingortext_embeddingtask, controls when text is split into chunks. When set totrue, a request from Elasticsearch contains only chunks related to a single document. Instead of batching chunks across documents, Elasticsearch sends them in separate requests. This ensures that chunk embeddings retain context from the entire document, improving semantic quality.If a document exceeds the model's context limits, or if the document contains non-text inputs (relevant when using the multimodal
embeddingtask), late chunking is automatically disabled for that document only and standard chunking is used instead.If not specified, defaults to
false. -
top_nnumber -
For a
reranktask, the number of most relevant documents to return. It defaults to the number of the documents. If this inference endpoint is used in atext_similarity_rerankerretriever query andtop_nis set, it must be greater than or equal torank_window_sizein the query.
Response
-
chunking_settingsInferenceChunkingSettings -
The chunking configuration object. Applies only to the
embedding,sparse_embeddingandtext_embeddingtask types. Not applicable to thererank,completion, orchat_completiontask types. -
max_chunk_sizenumber · default: 250 -
The maximum size of a chunk in words. This value cannot be lower than
20(forsentencestrategy) or10(forwordstrategy). This value should not exceed the window size for the associated model. -
overlapnumber · default: 100 -
The number of overlapping words for chunks. It is applicable only to a
wordchunking strategy. This value cannot be higher than half themax_chunk_sizevalue. -
sentence_overlapnumber · default: 1 -
The number of overlapping sentences for chunks. It is applicable only for a
sentencechunking strategy. It can be either1or0. -
separator_groupstring -
Only applicable to the
recursivestrategy and required when using it.Sets a predefined list of separators in the saved chunking settings based on the selected text type. Values can be
markdownorplaintext.Using this parameter is an alternative to manually specifying a custom
separatorslist. -
separatorsstring -
Only applicable to the
recursivestrategy and required when using it.A list of strings used as possible split points when chunking text.
Each string can be a plain string or a regular expression (regex) pattern. The system tries each separator in order to split the text, starting from the first item in the list.
After splitting, it attempts to recombine smaller pieces into larger chunks that stay within the
max_chunk_sizelimit, to reduce the total number of chunks generated. -
servicestringrequired -
The service type
-
service_settingsServiceSettingsrequired -
Settings specific to the service
-
task_settingsTaskSettings -
Task settings specific to the service and task type
-
inference_idstringrequired -
The inference Id