API Reference

Retrievers

Copy Markdown

Open in Claude

Open in ChatGPT

Open in Cursor

---

Copy Markdown

View as Markdown

Direct Retrieve

retrievers.search(RetrieverSearchParams**kwargs) -> CompositeRetrievalResult

POST/api/v1/retrievers/retrieve

Retrieve data using specified pipelines without creating a persistent retriever.

ParametersExpand Collapse

query: str

The query to retrieve against.

minLength1

organization_id: Optional[str]

project_id: Optional[str]

mode: Optional[CompositeRetrievalMode]

The mode of composite retrieval.

Accepts one of the following:

"routing"

"full"

pipelines: Optional[Iterable[RetrieverPipelineParam]]

The pipelines to use for retrieval.

description: Optional[str]

A description of the retriever tool.

maxLength15000

name: Optional[str]

A name for the retriever tool. Will default to the pipeline name if not provided.

maxLength3000

minLength1

pipeline_id: str

The ID of the pipeline this tool uses.

formatuuid

preset_retrieval_parameters: Optional[PresetRetrievalParams]

Parameters for retrieval configuration.

alpha: Optional[float]

Alpha value for hybrid retrieval to determine the weights between dense and sparse retrieval. 0 is sparse retrieval and 1 is dense retrieval.

maximum1

minimum0

class_name: Optional[str]

dense_similarity_cutoff: Optional[float]

Minimum similarity score wrt query for retrieval

maximum1

minimum0

dense_similarity_top_k: Optional[int]

Number of nodes for dense retrieval.

maximum100

minimum1

enable_reranking: Optional[bool]

Enable reranking for retrieval

files_top_k: Optional[int]

Number of files to retrieve (only for retrieval mode files_via_metadata and files_via_content).

maximum5

minimum1

rerank_top_n: Optional[int]

Number of reranked nodes for returning.

maximum100

minimum1

retrieval_mode: Optional[RetrievalMode]

The retrieval mode for the query.

Accepts one of the following:

"chunks"

"files_via_metadata"

"files_via_content"

"auto_routed"

Deprecatedretrieve_image_nodes: Optional[bool]

Whether to retrieve image nodes.

retrieve_page_figure_nodes: Optional[bool]

Whether to retrieve page figure nodes.

retrieve_page_screenshot_nodes: Optional[bool]

Whether to retrieve page screenshot nodes.

search_filters: Optional[MetadataFilters]

Metadata filters for vector stores.

filters: List[Filter]

Accepts one of the following:

class FilterMetadataFilter: …

Comprehensive metadata filter for vector stores to support more operators.

Value uses Strict types, as int, float and str are compatible types and were all converted to string before.

See: https://docs.pydantic.dev/latest/usage/types/#strict-types

key: str

value: Union[float, str, List[str], 3 more]

Accepts one of the following:

float

str

List[str]

List[float]

List[int]

operator: Optional[Literal["==", ">", "<", 11 more]]

Vector store filter operator.

Accepts one of the following:

"=="

">"

"<"

"!="

">="

"<="

"in"

"nin"

"any"

"all"

"text_match"

"text_match_insensitive"

"contains"

"is_empty"

class MetadataFilters: …

Metadata filters for vector stores.

filters: List[Filter]

Accepts one of the following:

class FilterMetadataFilter: …

Comprehensive metadata filter for vector stores to support more operators.

Value uses Strict types, as int, float and str are compatible types and were all converted to string before.

See: https://docs.pydantic.dev/latest/usage/types/#strict-types

key: str

value: Union[float, str, List[str], 3 more]

Accepts one of the following:

float

str

List[str]

List[float]

List[int]

operator: Optional[Literal["==", ">", "<", 11 more]]

Vector store filter operator.

Accepts one of the following:

"=="

">"

"<"

"!="

">="

"<="

"in"

"nin"

"any"

"all"

"text_match"

"text_match_insensitive"

"contains"

"is_empty"

MetadataFilters

condition: Optional[Literal["and", "or", "not"]]

Vector store filter conditions to combine different filters.

Accepts one of the following:

"and"

"or"

"not"

condition: Optional[Literal["and", "or", "not"]]

Vector store filter conditions to combine different filters.

Accepts one of the following:

"and"

"or"

"not"

search_filters_inference_schema: Optional[Dict[str, Union[Dict[str, object], List[object], str, 3 more]]]

JSON Schema that will be used to infer search_filters. Omit or leave as null to skip inference.

Accepts one of the following:

Dict[str, object]

List[object]

str

float

bool

sparse_similarity_top_k: Optional[int]

Number of nodes for sparse retrieval.

maximum100

minimum1

rerank_config: Optional[ReRankConfigParam]

The rerank configuration for composite retrieval.

top_n: Optional[int]

The number of nodes to retrieve after reranking over retrieved nodes from all retrieval tools.

minimum1

type: Optional[Literal["system_default", "llm", "cohere", 3 more]]

The type of reranker to use.

Accepts one of the following:

"system_default"

"llm"

"cohere"

"bedrock"

"score"

"disabled"

Deprecatedrerank_top_n: Optional[int]

(use rerank_config.top_n instead) The number of nodes to retrieve after reranking over retrieved nodes from all retrieval tools.

ReturnsExpand Collapse

class CompositeRetrievalResult: …

Deprecatedimage_nodes: Optional[List[PageScreenshotNodeWithScore]]

The image nodes retrieved by the pipeline for the given query. Deprecated - will soon be replaced with 'page_screenshot_nodes'.

node: Node

file_id: str

The ID of the file that the page screenshot was taken from

formatuuid

image_size: int

The size of the image in bytes

minimum0

page_index: int

The index of the page for which the screenshot is taken (0-indexed)

minimum0

metadata: Optional[Dict[str, object]]

Metadata for the screenshot

score: float

The score of the screenshot node

class_name: Optional[str]

nodes: Optional[List[Node]]

The retrieved nodes from the composite retrieval.

node: NodeNode

id: str

The ID of the retrieved node.

formatuuid

end_char_idx: Optional[int]

The end character index of the retrieved node in the document

pipeline_id: str

The ID of the pipeline this node was retrieved from.

formatuuid

retriever_id: str

The ID of the retriever this node was retrieved from.

formatuuid

retriever_pipeline_name: str

The name of the retrieval pipeline this node was retrieved from.

start_char_idx: Optional[int]

The start character index of the retrieved node in the document

text: str

The text of the retrieved node.

metadata: Optional[Dict[str, object]]

Metadata associated with the retrieved node.

class_name: Optional[str]

score: Optional[float]

page_figure_nodes: Optional[List[PageFigureNodeWithScore]]

The page figure nodes retrieved by the pipeline for the given query.

node: Node

confidence: float

The confidence of the figure

maximum1

minimum0

figure_name: str

The name of the figure

figure_size: int

The size of the figure in bytes

minimum0

file_id: str

The ID of the file that the figure was taken from

formatuuid

page_index: int

The index of the page for which the figure is taken (0-indexed)

minimum0

is_likely_noise: Optional[bool]

Whether the figure is likely to be noise

metadata: Optional[Dict[str, object]]

Metadata for the figure

score: float

The score of the figure node

class_name: Optional[str]

Direct Retrieve

Python

HTTP

TypeScript

Python

import os
from llama_cloud import LlamaCloud

client = LlamaCloud(
    api_key=os.environ.get("LLAMA_CLOUD_API_KEY"),  # This is the default and can be omitted
)
composite_retrieval_result = client.retrievers.search(
    query="x",
)
print(composite_retrieval_result.image_nodes)

200 example

{
  "image_nodes": [
    {
      "node": {
        "file_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
        "image_size": 0,
        "page_index": 0,
        "metadata": {
          "foo": "bar"
        }
      },
      "score": 0,
      "class_name": "class_name"
    }
  ],
  "nodes": [
    {
      "node": {
        "id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
        "end_char_idx": 0,
        "pipeline_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
        "retriever_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
        "retriever_pipeline_name": "retriever_pipeline_name",
        "start_char_idx": 0,
        "text": "text",
        "metadata": {
          "foo": "bar"
        }
      },
      "class_name": "class_name",
      "score": 0
    }
  ],
  "page_figure_nodes": [
    {
      "node": {
        "confidence": 0,
        "figure_name": "figure_name",
        "figure_size": 0,
        "file_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
        "page_index": 0,
        "is_likely_noise": true,
        "metadata": {
          "foo": "bar"
        }
      },
      "score": 0,
      "class_name": "class_name"
    }
  ]
}

Returns Examples

200 example

{
  "image_nodes": [
    {
      "node": {
        "file_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
        "image_size": 0,
        "page_index": 0,
        "metadata": {
          "foo": "bar"
        }
      },
      "score": 0,
      "class_name": "class_name"
    }
  ],
  "nodes": [
    {
      "node": {
        "id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
        "end_char_idx": 0,
        "pipeline_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
        "retriever_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
        "retriever_pipeline_name": "retriever_pipeline_name",
        "start_char_idx": 0,
        "text": "text",
        "metadata": {
          "foo": "bar"
        }
      },
      "class_name": "class_name",
      "score": 0
    }
  ],
  "page_figure_nodes": [
    {
      "node": {
        "confidence": 0,
        "figure_name": "figure_name",
        "figure_size": 0,
        "file_id": "182bd5e5-6e1a-4fe4-a799-aa6d9a6ab26e",
        "page_index": 0,
        "is_likely_noise": true,
        "metadata": {
          "foo": "bar"
        }
      },
      "score": 0,
      "class_name": "class_name"
    }
  ]
}