Multiple Corpora Query

POST 
https://api.vectara.io/v2/query

Perform a multipurpose query across to retrieve relevant information from one or more corpora and generate a response using Retrieval Augmented Generation (RAG).

• Specify the unique corpus_key identifying the corpus to query. The corpus_key is created in the Vectara Console UI or the Create Corpus API definition. When creating a new corpus, you have the option to assign a custom corpus_key following your preferred naming convention. This key serves as a unique identifier for the corpus, allowing it to be referenced in search requests. For more information, see Corpus Key Definition.
• Customize your search by specifying the query text (query), pagination details (offset and limit), and metadata filters (metadata_filter) to tailor your search results. Learn more
• Leverage advanced search capabilities like reranking (reranker) and opt-in Retrieval Augmented Generation (RAG) (generation) for enhanced query performance. Generation is opt in by setting the generation property. By excluding the property or by setting it to null, the response will not include generation. Learn more
• Specify Vectara's RAG-focused LLM (Mockingbird) for the generation_preset_name. Learn more
• Use advanced summarization options that utilize detailed summarization parameters such as max_response_characters, temperature, and frequency_penalty for generating precise and relevant summaries. Learn more
• Customize citation formats in summaries using the citations object to include numeric, HTML, or Markdown links. Learn more

For more detailed information, see this Query API guide.

Request

Header Parameters

Request-Timeout integer

Possible values: >= 1

The API will make a best effort to complete the request in the specified seconds or time out.

Request-Timeout-Millis integer

Possible values: >= 1

The API will make a best effort to complete the request in the specified milliseconds or time out.

application/json

Body

query stringrequired

The search query string, which is the question the user is asking.

search object

Search parameters to retrieve knowledge for the query.

corpora object[]

Possible values: >= 1

The corpora that you want to search.

Array [

custom_dimensions object

The custom dimensions as additional weights.

property name* double

metadata_filter string

The filter string used to narrow the search based on metadata attributes. The query against this corpus will be confined to document parts that match the metadata_filter. Only metadata fields set as filter_attributes on the corpus can be filtered. Filter syntax is similar to a SQL WHERE clause. See metadata filters documentation for more information.

lexical_interpolation float

Possible values: <= 1

How much to weigh lexical scores compared to the embedding score. 0 means lexical search is not used at all, and 1 means only lexical search is used.

semantics SearchSemantics

Possible values: [default, query, response]

Default value: default

Indicates whether to consider a query against this corpus as a query or a response.

corpus_key CorpusKeyrequired

Possible values: <= 50 characters, Value must match regular expression [a-zA-Z0-9_\=\-]+$

A user-provided key for a corpus.

query string

Query for a particular corpus that will override the overall query if present.

]

offset int32

Specifies how many results into the result to skip. This is useful for pagination.

limit int32

Possible values: >= 1

Default value: 10

The maximum number of results returned.

context_configuration object

Configuration on the presentation of each document part in the result set.

characters_before int32

The number of characters that are shown before the matching document part. This is useful to show the context of the document part in the wider document. Ignored if sentences_before is set. Vectara will capture the full sentence that contains the captured characters, to not lose the meaning caused by a truncated word or sentence.

characters_after int32

The number of characters that are shown after the matching document part. This is useful to show the context of the document part in the wider document. Ignored if sentences_after is set. Vectara will capture the full sentence that contains the captured characters, to not lose the meaning caused by a truncated word or sentence.

sentences_before int32

The number of sentences that are shown before the matching document part. This is useful to show the context of the document part in the wider document.

sentences_after int32

The number of sentences that are shown after the matching document part. This is useful to show the context of the document part in the wider document.

start_tag string

The tag that wraps the document part at the start. This is often used to provide a start HTML/XML tag or some other delimiter you can use in an application to understand where to provide highlighting in your UI and understand where the context before ends and the document part begins.

end_tag string

The tag that wraps the document part at the end. This is often used to provide a start HTML/XML tag or some other delimiter you can use in an application to understand where to provide highlighting in your UI and understand where the context before ends and the document part begins.

reranker object

Rerank results of the search. Rerankers are very powerful tools to improve the order of search results. By default the search will use the most powerful reranker available to the customer's plan. To disable reranking, set the reranker type to "none".

oneOf

CustomerSpecificReranker

type string

Default value: customer_reranker

When the type is customer_reranker, you can specify the reranker_name of a reranker. reranker_id is deprecated. The retrieval engine will then rerank results using that reranker.

reranker_id stringdeprecated

Possible values: Value must match regular expression rnk_(?!272725718)\d+

The ID of the reranker. The multilingual reranker that may be specified is rnk_272725719. Do not specify the MMR reranker ID here, and instead, use the MMR reranker object type. Deprecated: Use reranker_name instead.

reranker_name string

The name of the reranker. Do not specify the MMR reranker name here. Instead, use the MMR reranker object type.

limit int32

Possible values: >= 1

Specifies the maximum number of results to be returned after the reranking process. When a reranker is applied, it performs the following steps:

1. Reranks all input results according to its algorithm.
2. Sorts the reranked results based on their new scores.
3. Returns the top N results, where N is the value specified by this limit.

Note: This limit is applied per reranking stage. In a chain of rerankers, each reranker can have its own limit, potentially reducing the number of results at each stage.

cutoff float

Specifies the minimum score threshold for results to be included after the reranking process. When a reranker is applied with a cutoff, it performs the following steps:

1. Reranks all input results according to its algorithm.
2. Applies the cutoff, removing any results with scores below the specified threshold.
3. Returns the remaining results, sorted by their new scores.

Note: This cutoff is applied per reranking stage. In a chain of rerankers, each reranker can have its own cutoff, potentially further reducing the number of results at each stage. If both 'limit' and 'cutoff' are specified, the cutoff is applied first, followed by the limit.

UserFunctionReranker

type string

Default value: userfn

When the type is userfn, you can define custom reranking functions using document-level metadata, part-level metadata, or scores generated from the request-level metadata.

user_function string

The user defined function.

limit int32

Possible values: >= 1

Specifies the maximum number of results to be returned after the reranking process. When a reranker is applied, it performs the following steps:

1. Reranks all input results according to its algorithm.
2. Sorts the reranked results based on their new scores.
3. Returns the top N results, where N is the value specified by this limit.

Note: This limit is applied per reranking stage. In a chain of rerankers, each reranker can have its own limit, potentially reducing the number of results at each stage.

cutoff float

Specifies the minimum score threshold for results to be included after the reranking process. When a reranker is applied with a cutoff, it performs the following steps:

1. Reranks all input results according to its algorithm.
2. Applies the cutoff, removing any results with scores below the specified threshold.
3. Returns the remaining results, sorted by their new scores.

Note: This cutoff is applied per reranking stage. In a chain of rerankers, each reranker can have its own cutoff, potentially further reducing the number of results at each stage. If both 'limit' and 'cutoff' are specified, the cutoff is applied first, followed by the limit.

MMRReranker

type string

Default value: mmr

When the type is mmr, you can specify the diversity_bias, and the retrieval engine will use the MMR reranker.

diversity_bias float

The diversity bias. Higher values indicate more diversity.

limit int32

Possible values: >= 1

Specifies the maximum number of results to be returned after the reranking process. When a reranker is applied, it performs the following steps:

1. Reranks all input results according to its algorithm.
2. Sorts the reranked results based on their new scores.
3. Returns the top N results, where N is the value specified by this limit.

Note: This limit is applied per reranking stage. In a chain of rerankers, each reranker can have its own limit, potentially reducing the number of results at each stage.

cutoff float

Specifies the minimum score threshold for results to be included after the reranking process. When a reranker is applied with a cutoff, it performs the following steps:

1. Reranks all input results according to its algorithm.
2. Applies the cutoff, removing any results with scores below the specified threshold.
3. Returns the remaining results, sorted by their new scores.

Note: This cutoff is applied per reranking stage. In a chain of rerankers, each reranker can have its own cutoff, potentially further reducing the number of results at each stage. If both 'limit' and 'cutoff' are specified, the cutoff is applied first, followed by the limit.

ChainReranker

type string

Default value: chain

When the type is chain, you can then chain re-rankers together.

rerankers arrayrequired

Possible values: <= 50

Specify an array of rerankers to apply to search results consecutively.

NoneReranker

type string

Default value: none

When the type is none, no reranking will be done.

limit int32

Possible values: >= 1

Specifies the maximum number of results to be returned after the reranking process. When a reranker is applied, it performs the following steps:

1. Reranks all input results according to its algorithm.
2. Sorts the reranked results based on their new scores.
3. Returns the top N results, where N is the value specified by this limit.

Note: This limit is applied per reranking stage. In a chain of rerankers, each reranker can have its own limit, potentially reducing the number of results at each stage.

generation object

The parameters to control generation.

generation_preset_name string

Possible values: non-empty

The preset values to use to feed the query results and other context to the model.

A generation_preset is an object with a bundle of properties that specifies:

• The prompt_template that is rendered and then sent to the LLM.
• The LLM used.
• model_parameters such as temperature.

All of these properties except the model can be overridden by setting them in this object. Even when a prompt_template is set, the generation_preset_name is used to set the model used. See model_parameters.model if you want to set the model explicitly.

If generation_preset_name is not set, the Vectara platform will use the default model and prompt.

prompt_name stringdeprecated

Possible values: non-empty

Use generation_preset_name instead of prompt_name.

max_used_search_results int32

Default value: 5

The maximum number of search results to be available to the prompt.

prompt_template string

Vectara manages both system and user roles and prompts for the generative LLM out of the box by default. However, users can override the prompt_template via this variable. The prompt_template is in the form of an Apache Velocity template. For more details on how to configure the prompt_template, see the long-form documentation.

prompt_text stringdeprecated

This property is deprecated in favor of clearer naming. Use prompt_template. This property will be ignored if prompt_template is set.

max_response_characters int32

Controls the length of the generated output. This is a rough estimate and not a hard limit: the end output can be longer or shorter than this value. This is generally implemented by including the max_response_characters in the prompt, and the LLM's instruction following capability dictates how closely the generated output is limited.

response_language Language

Possible values: [auto, eng, deu, fra, zho, kor, ara, rus, tha, nld, ita, por, spa, jpn, pol, tur, vie, ind, ces, ukr, ell, heb, fas, hin, urd, swe, ben, msa, ron]

Default value: auto

Languages that the Vectara platform supports.

model_parameters object

The parameters for the model. WARNING: This is an experimental feature, and breakable at any point with virtually no notice. It is meant for experimentation to converge on optimal parameters that can then be set in the prompt definitions.

llm_name string

The model (e.g., gpt-4) to use for summarization. If specified, it will override the model behind generation_preset_name.

max_tokens int32

Possible values: >= 1

The maximum number of tokens to be returned by the model.

temperature float

The sampling temperature to use. Higher values make the output more random, while lower values make it more focused and deterministic.

frequency_penalty float

Higher 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.

presence_penalty float

Higher values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics.

citations object

Style the generator should use when making citations.

style string

Possible values: [none, numeric, html, markdown]

The citation style to be used in summary. Can be one of:

• numeric - Citations formatted as simple numerals: [1], [2] ...
• none - Citations removed from text.
• html - Citation formatted as a URL like <a href="url_pattern">text_pattern</a>.
• markdown - Formatted as [text_pattern](url_pattern).

url_pattern string

The URL pattern if the citation_style is set to html or markdown. The pattern can access metadata attributes in the document or part. e.g. https://my.doc/foo/{doc.id}/{part.id}

The default url_pattern is an empty string.

text_pattern string

The text pattern if the citation_style is set to html or markdown. This pattern sets the href for HTML or the text within [] in markdown, and defaults to N being the index of result if it is not set.

The default citation style looks like [N](<url_pattern>) for markdown.

You can use metadata attributes in the text_pattern. For example, the pattern {doc.title} with citation style markdown would result in final citation output like [Title](<url_pattern>) when the document's metadata includes {"title":"Title"}.

enable_factual_consistency_score boolean

Default value: true

Enable returning the factual consistency score with query results.

stream_response boolean

Default value: false

Indicates whether the response should be streamed or not.

save_history boolean

Indicates whether to save the query to query history.

intelligent_query_rewriting boolean

Default value: false

Indicates whether to enable intelligent query rewriting. When enabled, the platform will attempt to extract metadata filter and rewrite the query to improve search results.

Responses

200

A response to a query.

application/json

Schema

Schema

summary string

The summary of the search results.

response_language Language

Possible values: [auto, eng, deu, fra, zho, kor, ara, rus, tha, nld, ita, por, spa, jpn, pol, tur, vie, ind, ces, ukr, ell, heb, fas, hin, urd, swe, ben, msa, ron]

Default value: auto

Languages that the Vectara platform supports.

search_results object[]
The ranked search results.
Array [
text string
The document part altered by the context configuration that matches the query.
score double
The score of the individual result.
part_metadata object
The metadata for the document part.
property name* any
The metadata for the document part.
document_metadata object
The metadata for the document that contains the document part.
property name* any
The metadata for the document that contains the document part.
document_id string
The ID of the document that contains the document part.
table object
The table that the document part is from.
id string
The unique ID of the table within the document.
title string
The title of the table.
data object
The data of the table.
headers array[]
The headers of the table.
rows array[]
The rows in the data.
description string
The description of the table.
request_corpora_index int32
A query request can search over multiple corpora at a time. This property is set to the index in the list of corpora in the original search request that this search result originated from.
If the query request is only over one corpus, this property is 0.
]
factual_consistency_score float

Indicates the probability that the summary is factually consistent with the results. The system excludes this property if it encounters excessively large outputs or search results.

rendered_prompt string

The rendered prompt sent to the LLM. Useful when creating customer prompt_template templates.

warnings QueryWarning[]

Possible values: [exceeded_max_input_length_fcs, intelligent_query_rewriting_failed]

Non-fatal warnings that occurred during request processing

rewritten_queries object[]
The rewritten queries for the corpora that were searched. Only populated when intelligent_query_rewriting is enabled.
Array [
corpus_key string
The corpus key that the query was made on.
filter_extraction object
The result of query filter extraction.
query string
The query rephrased from the input query and executed.
metadata_filter string
The metadata filter extracted from the input query.
]

Example (from schema)

{
  "summary": "string",
  "response_language": "auto",
  "search_results": [
    {
      "text": "string",
      "score": 0,
      "part_metadata": {},
      "document_metadata": {},
      "document_id": "string",
      "table": {
        "id": "table_1",
        "title": "string",
        "data": {
          "headers": [
            [
              {
                "text_value": "string",
                "int_value": 0,
                "float_value": 0,
                "bool_value": true,
                "colspan": 0,
                "rowspan": 0
              }
            ]
          ],
          "rows": [
            [
              {
                "text_value": "string",
                "int_value": 0,
                "float_value": 0,
                "bool_value": true,
                "colspan": 0,
                "rowspan": 0
              }
            ]
          ]
        },
        "description": "string"
      },
      "request_corpora_index": 0
    }
  ],
  "factual_consistency_score": 0,
  "rendered_prompt": "string",
  "warnings": [
    "exceeded_max_input_length_fcs"
  ],
  "rewritten_queries": [
    {
      "corpus_key": "string",
      "filter_extraction": {
        "query": "string",
        "metadata_filter": "string"
      }
    }
  ]
}

text/event-stream

Schema

Schema

oneOf

StreamSearchResponse

type string

Default value: search_results

When the streaming event has the search results, the type will be search_results.

search_results object[]

The ranked search results.

Array [

text string

The document part altered by the context configuration that matches the query.

score double

The score of the individual result.

part_metadata object

The metadata for the document part.

property name* any

The metadata for the document part.

document_metadata object

The metadata for the document that contains the document part.

property name* any

The metadata for the document that contains the document part.

document_id string

The ID of the document that contains the document part.

table object

The table that the document part is from.

id string

The unique ID of the table within the document.

title string

The title of the table.

data object

The data of the table.

headers array[]

The headers of the table.

rows array[]

The rows in the data.

description string

The description of the table.

request_corpora_index int32

A query request can search over multiple corpora at a time. This property is set to the index in the list of corpora in the original search request that this search result originated from.

If the query request is only over one corpus, this property is 0.

]

rewritten_queries object[]

The rewritten queries for the corpora that were searched. Only populated when intelligent_query_rewriting is enabled.

Array [

corpus_key string

The corpus key that the query was made on.

filter_extraction object

The result of query filter extraction.

query string

The query rephrased from the input query and executed.

metadata_filter string

The metadata filter extracted from the input query.

]

StreamGenerationChunk

type string

Default value: generation_chunk

When the streaming event contains the next chunk of generator output, the type will be generation_chunk.

generation_chunk string

Part of the message from the generator. All summary chunks must be appended together in order to get the full summary.

StreamGenerationEnd

type string

Default value: generation_end

Then end of generation will be denoted with an object with the type generation_end.

StreamResponseEnd

type string

Default value: end

Then end of stream will be denoted with an object with the type end.

FactualConsistencyScore

type string

Default value: factual_consistency_score

When the streaming event contains the factual consistency score, the type will be factual_consistency_score.

factual_consistency_score float

The probability that the summary is factually consistent with the results.

GenerationInfo

type string

Default value: generation_info

When the streaming event contains the generation information type will be generation_info.

rendered_prompt string

The rendered prompt sent to the LLM. Useful when creating customer prompt_template templates.

rephrased_query string

View the actual query made to backend that was rephrased by the LLM from the input query.

StreamError

type string

Default value: error

If the stream errors, an event with type error will be sent.

messages string[]

The error messages.

Example (from schema)

{}

400

Query request was malformed.

application/json

Schema

Schema

field_errors object
The errors that relate to specific fields in the request.
property name* string
messages message[]
request_id string

The ID of the request that can be used to help Vectara support debug what went wrong.

Example (from schema)

{
  "field_errors": {},
  "messages": [
    "string"
  ],
  "request_id": "string"
}

403

Permissions do not allow querying one or more corpora in the request.

application/json

Schema

Schema

messages message[]

The messages describing why the error occurred.

request_id string

The ID of the request that can be used to help Vectara support debug what went wrong.

Example (from schema)

{
  "messages": [
    "Internal server error."
  ],
  "request_id": "string"
}

404

One or more of the corpora were not found.

application/json

Schema

Schema

id string

The ID cannot be found.

messages message[]
request_id string

ID of the request that can be used to help Vectara support debug what went wrong.

Example (from schema)

{
  "id": "string",
  "messages": [
    "string"
  ],
  "request_id": "string"
}

Authorization: x-api-key

name: x-api-keytype: apiKeyin: header

• curl
• python
• go
• nodejs
• ruby
• csharp
• php
• java
• powershell

• CURL

curl -L -X POST 'https://api.vectara.io/v2/query' \
-H 'Content-Type: application/json' \
-H 'Accept: application/json' \
-H 'x-api-key: <API_KEY_VALUE>' \
--data-raw '{
  "query": "Am I allowed to bring pets to work?",
  "search": {
    "corpora": [
      {
        "custom_dimensions": {},
        "metadata_filter": "doc.title = '\''Charlotte'\'''\''s Web'\''",
        "lexical_interpolation": 0.025,
        "semantics": "default",
        "corpus_key": "my-corpus",
        "query": "string"
      }
    ],
    "offset": 0,
    "limit": 10,
    "context_configuration": {
      "characters_before": 30,
      "characters_after": 30,
      "sentences_before": 3,
      "sentences_after": 3,
      "start_tag": "<em>",
      "end_tag": "</em>"
    },
    "reranker": {
      "type": "customer_reranker",
      "reranker_name": "Rerank_Multilingual_v1",
      "limit": 0,
      "cutoff": 0
    }
  },
  "generation": {
    "generation_preset_name": "vectara-summary-ext-v1.2.0",
    "max_used_search_results": 5,
    "prompt_template": "[\n  {\"role\": \"system\", \"content\": \"You are a helpful search assistant.\"},\n  #foreach ($qResult in $vectaraQueryResults)\n     {\"role\": \"user\", \"content\": \"Given the $vectaraIdxWord[$foreach.index] search result.\"},\n     {\"role\": \"assistant\", \"content\": \"${qResult.getText()}\" },\n  #end\n  {\"role\": \"user\", \"content\": \"Generate a summary for the query '\''${vectaraQuery}'\'' based on the above results.\"}\n]\n",
    "max_response_characters": 300,
    "response_language": "auto",
    "model_parameters": {
      "llm_name": "gpt4",
      "max_tokens": 0,
      "temperature": 0,
      "frequency_penalty": 0,
      "presence_penalty": 0
    },
    "citations": {
      "style": "none",
      "url_pattern": "https://vectara.com/documents/{doc.id}",
      "text_pattern": "{doc.title}"
    },
    "enable_factual_consistency_score": true
  },
  "stream_response": false,
  "save_history": true,
  "intelligent_query_rewriting": false
}'

Request Collapse all

Base URL

https://api.vectara.io

Auth

Security Scheme

ApiKeyAuthOAuth2

ApiKeyAuth

Parameters

Request-Timeout — header

Request-Timeout-Millis — header

Body

{
  "query": "Am I allowed to bring pets to work?",
  "search": {
    "corpora": [
      {
        "custom_dimensions": {},
        "metadata_filter": "doc.title = 'Charlotte''s Web'",
        "lexical_interpolation": 0.025,
        "semantics": "default",
        "corpus_key": "my-corpus",
        "query": "string"
      }
    ],
    "offset": 0,
    "limit": 10,
    "context_configuration": {
      "characters_before": 30,
      "characters_after": 30,
      "sentences_before": 3,
      "sentences_after": 3,
      "start_tag": "<em>",
      "end_tag": "</em>"
    },
    "reranker": {
      "type": "customer_reranker",
      "reranker_name": "Rerank_Multilingual_v1",
      "limit": 0,
      "cutoff": 0
    }
  },
  "generation": {
    "generation_preset_name": "vectara-summary-ext-v1.2.0",
    "max_used_search_results": 5,
    "prompt_template": "[\n  {\"role\": \"system\", \"content\": \"You are a helpful search assistant.\"},\n  #foreach ($qResult in $vectaraQueryResults)\n     {\"role\": \"user\", \"content\": \"Given the $vectaraIdxWord[$foreach.index] search result.\"},\n     {\"role\": \"assistant\", \"content\": \"${qResult.getText()}\" },\n  #end\n  {\"role\": \"user\", \"content\": \"Generate a summary for the query '${vectaraQuery}' based on the above results.\"}\n]\n",
    "max_response_characters": 300,
    "response_language": "auto",
    "model_parameters": {
      "llm_name": "gpt4",
      "max_tokens": 0,
      "temperature": 0,
      "frequency_penalty": 0,
      "presence_penalty": 0
    },
    "citations": {
      "style": "none",
      "url_pattern": "https://vectara.com/documents/{doc.id}",
      "text_pattern": "{doc.title}"
    },
    "enable_factual_consistency_score": true
  },
  "stream_response": false,
  "save_history": true,
  "intelligent_query_rewriting": false
}

ResponseClear

Click the Send API Request button above and see the response here!