> ## Documentation Index
> Fetch the complete documentation index at: https://developer.box.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Generate text
> Sends an AI request to supported Large Language Models (LLMs) and returns generated text based on the provided prompt.
This endpoint is in the version **2024.0**. No changes are required to continue using it. For more details, see **[Box API versioning](/guides/api-calls/api-versioning-strategy)**.
Learn more about [Box SDK versioning strategy](/guides/tooling/sdks/sdk-versioning/).
## OpenAPI
````yaml /box-openapi.json POST /ai/text_gen
openapi: 3.0.2
info:
title: Box Platform API
description: >-
[Box Platform](https://developer.box.com) provides functionality to provide
access to content stored within [Box](https://box.com). It provides
endpoints for basic manipulation of files and folders, management of users
within an enterprise, as well as more complex topics such as legal holds and
retention policies.
termsOfService: https://cloud.app.box.com/s/rmwxu64h1ipr41u49w3bbuvbsa29wku9
contact:
name: Box, Inc
url: https://developer.box.com
email: devrel@box.com
license:
name: Apache-2.0
url: https://www.apache.org/licenses/LICENSE-2.0
version: '2024.0'
x-box-commit-hash: 95117b97f9
servers:
- url: https://api.box.com/2.0
description: Box Platform API server.
security:
- OAuth2Security: []
tags:
- name: AI
description: A set of endpoints used to interact with supported LLMs.
x-box-tag: ai
- name: AI Studio
description: A set of endpoints used to interact with AI Studio.
x-box-tag: ai_studio
- name: App item associations
x-box-tag: app_item_associations
- name: Authorization
description: A set of endpoints used to manage user authorization process.
x-box-tag: authorization
x-box-priority: true
- name: Box Sign requests
description: Box Sign requests are used to submit a file for signature.
x-box-tag: sign_requests
- name: Classifications
description: >-
Classification labels are used for content that is sensitive or under
security restrictions.
x-box-tag: classifications
- name: Classifications on files
description: >-
Classification labels are used for files that are sensitive or under
security restrictions.
x-box-tag: file_classifications
- name: Classifications on folders
description: >-
Classification labels are used for folders that are sensitive or under
security restrictions.
x-box-tag: folder_classifications
- name: Collaborations
description: >-
Collaborations define access permissions for users and groups to files and
folders, similar to access control lists.
x-box-tag: user_collaborations
- name: Collaborations (List)
description: >-
A set of endpoints used to retrieve file, folder, pending, and group
collaborations.
x-box-tag: list_collaborations
- name: Collections
description: >-
Collections are a way to group files, folders, and web links without
putting them all into a folder.
x-box-tag: collections
- name: Comments
description: >-
Comments are messages generated users on files, allowing users to
collaborate on a file, discussing any feedback they might have on the
content.
x-box-tag: comments
- name: Device pinners
description: >-
Device pinners allow enterprises to control what devices can use native
Box applications.
x-box-tag: device_pinners
- name: Domain restrictions (User exemptions)
description: >-
A set of endpoints that allow exempting users from restrictions imposed by
the list of allowed collaboration domains for a specific enterprise.
x-box-tag: collaboration_allowlist_exempt_targets
- name: Domain restrictions for collaborations
description: >-
A set of endpoints that manage domains for which users can collaborate
with files and folders in an enterprise.
x-box-tag: collaboration_allowlist_entries
- name: Downloads
description: >-
Downloads allow saving files to the application's server, or directly by
the end user in a browser.
x-box-tag: downloads
- name: Email aliases
description: >-
Email aliases provide a list of emails additional to the user's primary
login email.
x-box-tag: email_aliases
- name: Events
description: >-
Events provide a way for an application to subscribe to any actions
performed by any user, users, or service in an enterprise.
x-box-tag: events
- name: File requests
description: >-
File Requests provide a fast and secure way to request files and
associated metadata from anyone. Users can create new file requests based
on an existing file request, update file request settings, activate,
deactivate, and delete file requests programmatically.
x-box-tag: file_requests
- name: File version legal holds
description: >-
A legal hold is a process that an enterprise can use to preserve all forms
of potentially relevant information when litigation is pending or
reasonably anticipated. A File Version Legal Hold represents all the
policies that are assigned to a specific file version.
x-box-tag: file_version_legal_holds
- name: File version retentions
description: >-
A retention policy blocks permanent deletion of content for a specified
amount of time. A file version retention is a record for a retained file.
x-box-tag: file_version_retentions
- name: File versions
description: A set of endpoints used to manage specific versions of a file.
x-box-tag: file_versions
- name: Files
description: >-
Files, together with Folders, are at the core of the Box API. Files can be
uploaded and downloaded, as well as hold important metadata information
about the content.
x-box-tag: files
- name: Folder Locks
description: >-
Folder locks define access restrictions placed by folder owners to prevent
specific folders from being moved or deleted.
x-box-tag: folder_locks
- name: Folders
description: >-
Folders, together with Files, are at the core of the Box API. Folders can
be uploaded and downloaded, as well as hold important metadata information
about the content.
x-box-tag: folders
- name: Integration mappings
description: >-
Integration Mappings allow the users to manage where content from partner
apps is stored in Box.
x-box-tag: integration_mappings
- name: Group memberships
description: Group memberships signify that a user is a part of the group.
x-box-tag: memberships
- name: Groups
description: Groups created in an enterprise.
x-box-tag: groups
- name: Invites
description: Invites are used to invite the user to an enterprise.
x-box-tag: invites
- name: Legal hold policies
description: >-
A legal hold is a process that an enterprise can use to preserve all forms
of potentially relevant information when litigation is pending or
reasonably anticipated.
x-box-tag: legal_hold_policies
- name: Legal hold policy assignments
description: >-
A Legal Hold Policy Assignment is a relation between a policy and
custodian. In this case, as custodian can be a user, folder, file, or file
version.
x-box-tag: legal_hold_policy_assignments
- name: Metadata cascade policies
description: >-
A metadata cascade policy describes how metadata instances applied to a
folder should be applied to any item within that folder.
x-box-tag: metadata_cascade_policies
- name: Metadata instances (Files)
description: >-
A metadata instance describes the relation between a template and a file,
including the values that are assigned for every field.
x-box-tag: file_metadata
- name: Metadata instances (Folders)
description: >-
A metadata instance describes the relation between a template and a
folder, including the values that are assigned for every field.
x-box-tag: folder_metadata
- name: Metadata taxonomies
description: >-
A metadata taxonomy is a hierarchical classification system that helps
organize and manage metadata within an enterprise.
x-box-tag: metadata_taxonomies
- name: Metadata templates
description: >-
A metadata template describes a reusable set of key/value pairs that can
be assigned to a file.
x-box-tag: metadata_templates
- name: Recent items
description: >-
Recent items represent items such as files or folders that the user
accessed recently.
x-box-tag: recent_items
- name: Retention policies
description: >-
A retention policy blocks permanent deletion of content for a specified
amount of time. Admins can create retention policies and then assign them
to specific folders or their entire enterprise.
x-box-tag: retention_policies
- name: Retention policy assignments
description: >-
A Retention Policy Assignment is a relation between a policy and folder or
enterprise. Creating an assignment puts a retention on all the file
versions that belong to that folder or enterprise.
x-box-tag: retention_policy_assignments
- name: Search
description: >-
The Box API provides a way to find content in Box using full-text search
queries.
x-box-tag: search
- name: Session termination
description: >-
Session termination API is used to validate the roles and permissions of
the group, and creates asynchronous jobs to terminate the group's
sessions.
x-box-tag: session_termination
- name: Shared links (Files)
description: >-
Files shared links are URLs that are generated for files stored in Box,
which provide direct, read-only access to the resource.
x-box-tag: shared_links_files
- name: Shared links (Folders)
description: >-
Folders shared links are URLs that are generated for folders stored in
Box, which provide direct, read-only access to the resource.
x-box-tag: shared_links_folders
- name: Shared links (Web Links)
description: >-
Web links for files are URLs that are generated for web links in Box,
which provide direct, read-only access to the resource.
x-box-tag: shared_links_web_links
- name: Shared links (App Items)
description: >-
URLs generated for app items stored in Box, which provide direct,
read-only access to the resource.
x-box-tag: shared_links_app_items
- name: Shield information barriers
description: >-
Shield information barrier in Box defines an ethical wall. An ethical wall
is a mechanism that prevents exchanges or communication that could lead to
conflicts of interest and therefore result in business activities
ethically or legally questionable.
x-box-tag: shield_information_barriers
- name: Shield information barrier segments
description: >-
Shield information barrier segment represents a defined group of users. A
user can be a member of only one segment, which makes segments different
from groups.
x-box-tag: shield_information_barrier_segments
- name: Shield information barrier segment members
description: >-
Shield information barrier segment member represents a user that is
assigned to a specific segment.
x-box-tag: shield_information_barrier_segment_members
- name: Shield information barrier reports
description: >-
Shield information barrier reports contain information on what existing
collaborations will be removed permanently when the information barrier is
enabled.
x-box-tag: shield_information_barrier_reports
- name: Shield information barrier segment restrictions
description: >-
Shield information barrier segment restriction is an access restriction
based on the content (file or folder) owner.
x-box-tag: shield_information_barrier_segment_restrictions
- name: Box Sign templates
description: >-
Sign templates allow you to use a predefined Box Sign template when
creating a signature request. The template includes placeholders that are
automatically populated with data when creating the request.
x-box-tag: sign_templates
- name: Skills
description: >-
Box Skills are designed to allow custom processing of files uploaded to
Box, with the intent of enhancing the underlying metadata of the file.
x-box-tag: skills
- name: Standard and Zones Storage Policies
description: >-
Storage policy assignment represents the storage zone for items in a given
enterprise.
x-box-tag: storage_policies
- name: Standard and Zones Storage Policy Assignments
description: >-
Storage policy assignment represents the relation between storage zone and
the assigned item (for example a file stored in a specific zone).
x-box-tag: storage_policy_assignments
- name: Task assignments
description: >-
A task assignment defines which task is assigned to which user to
complete.
x-box-tag: task_assignments
- name: Tasks
description: >-
Tasks allow users to request collaborators on a file to review a file or
complete a piece of work. Tasks can be used by developers to create
file-centric workflows.
x-box-tag: tasks
- name: Terms of service
description: A set of endpoints used to manage terms of service agreements.
x-box-tag: terms_of_services
- name: Terms of service user statuses
description: >-
A set of endpoints used to manage the status of terms of service for a
particular user.
x-box-tag: terms_of_service_user_statuses
- name: Transfer folders
description: >-
API designed to move all of the items (files, folders and workflows) owned
by a user into another user's account.
x-box-tag: transfer
- name: Trashed files
description: Files that were deleted and are in trash.
x-box-tag: trashed_files
- name: Trashed folders
description: Folders that were deleted and are in trash.
x-box-tag: trashed_folders
- name: Trashed items
description: Items that were deleted and are in trash.
x-box-tag: trashed_items
- name: Trashed web links
description: Web links that were deleted and are in trash.
x-box-tag: trashed_web_links
- name: Uploads
description: >-
The direct file upload API supports files up to 50MB in size and sends all
the binary data to the Box API in 1 API request.
x-box-tag: uploads
- name: Uploads (Chunked)
description: >-
The chunked upload endpoints support files from 20MB in size and allow an
application to upload the file in parts, allowing for more control to
catch any errors and retry parts individually.
x-box-tag: chunked_uploads
- name: User avatars
description: >-
User avatars are JPG or PNG files uploaded to Box to represent the user
image. They are then displayed in the user account.
x-box-tag: avatars
- name: Users
description: >-
Box API supports a variety of users, ranging from real employees logging
in with their Managed User account, to applications using App Users to
drive powerful automation workflows.
x-box-tag: users
- name: Watermarks (Files)
description: >-
A watermark is a semi-transparent overlay on an embedded file preview that
displays a viewer's email address or user ID and the time of access over
the file.
x-box-tag: file_watermarks
- name: Watermarks (Folders)
description: >-
A watermark is a semi-transparent overlay on an embedded folder preview
that displays a viewer's email address or user ID and the time of access
over the folder content.
x-box-tag: folder_watermarks
- name: Web links
description: >-
Web links are objects that point to URLs. These objects are also known as
bookmarks within the Box web application.
x-box-tag: web_links
- name: Webhooks
description: >-
Webhooks allow you to monitor Box content for events, and receive
notifications to a URL of your choice when they occur. For example, a
workflow may include waiting for a file to be downloaded to delete a
shared link.
x-box-tag: webhooks
- name: Workflows
description: >-
Box Relay Workflows are objects that represent a named collection of
flows.
x-box-tag: workflows
- name: Zip Downloads
description: >-
Zip downloads represent a successful request to create a ZIP archive with
files and folders.
x-box-tag: zip_downloads
externalDocs:
description: Box Developer Documentation.
url: https://developer.box.com
paths:
/ai/text_gen:
post:
tags:
- AI
summary: Generate text
description: >-
Sends an AI request to supported Large Language Models (LLMs) and
returns generated text based on the provided prompt.
operationId: post_ai_text_gen
requestBody:
content:
application/json:
schema:
$ref: '#/components/schemas/AiTextGen'
responses:
'200':
description: A successful response including the answer from the LLM.
content:
application/json:
schema:
$ref: '#/components/schemas/AiResponse'
'500':
description: An unexpected server error.
content:
application/json:
schema:
$ref: '#/components/schemas/ClientError'
default:
description: An unexpected error.
content:
application/json:
schema:
$ref: '#/components/schemas/ClientError'
x-codeSamples:
- lang: curl
label: Generate text
source: |-
curl -i -L POST "https://api.box.com/2.0/ai/text_gen" \
-H "content-type: application/json" \
-H "authorization: Bearer " \
-d '{
"prompt": "Write a social media post about protein powder.",
"items": [
{
"id": "12345678",
"type": "file",
"content": "More information about protein powders"
},
],
"dialogue_history": [
{
"prompt": "Can you add some more information?",
"answer": "Public API schemas provide necessary information to integrate with APIs...",
"created_at": "2013-12-12T11:20:43-08:00"
}
],
"ai_agent": {
"type": "ai_agent_text_gen",
"basic_gen": {
"model": "azure__openai__gpt_4o_mini"
}
}
}'
- lang: dotnet
label: Generate text
source: >-
await client.Ai.CreateAiTextGenAsync(requestBody: new
AiTextGen(prompt: "Paraphrase the documents", items:
Array.AsReadOnly(new [] {new AiTextGenItemsField(id: fileToAsk.Id,
type: AiTextGenItemsTypeField.File) { Content = "The Earth goes
around the Sun. The Sun rises in the east in the morning." }})) {
DialogueHistory = Array.AsReadOnly(new [] {new AiDialogueHistory() {
Prompt = "What does the earth go around?", Answer = "The Sun",
CreatedAt = Utils.DateTimeFromString(dateTime:
"2021-01-01T00:00:00Z") },new AiDialogueHistory() { Prompt = "On
Earth, where does the Sun rise?", Answer = "east", CreatedAt =
Utils.DateTimeFromString(dateTime: "2021-01-01T00:00:00Z") }}) });
- lang: swift
label: Generate text
source: >-
try await client.ai.createAiTextGen(requestBody: AiTextGen(prompt:
"Paraphrase the documents", items: [AiTextGenItemsField(id:
fileToAsk.id, type: AiTextGenItemsTypeField.file, content: "The
Earth goes around the Sun. The Sun rises in the east in the
morning.")], dialogueHistory: [AiDialogueHistory(prompt: "What does
the earth go around?", answer: "The Sun", createdAt: try
Utils.Dates.dateTimeFromString(dateTime: "2021-01-01T00:00:00Z")),
AiDialogueHistory(prompt: "On Earth, where does the Sun rise?",
answer: "east", createdAt: try
Utils.Dates.dateTimeFromString(dateTime: "2021-01-01T00:00:00Z"))]))
- lang: java
label: Generate text
source: >-
client.getAi().createAiTextGen(new AiTextGen.Builder("Paraphrase the
documents", Arrays.asList(new
AiTextGenItemsField.Builder(fileToAsk.getId()).type(AiTextGenItemsTypeField.FILE).content("The
Earth goes around the Sun. The Sun rises in the east in the
morning.").build())).dialogueHistory(Arrays.asList(new
AiDialogueHistory.Builder().prompt("What does the earth go
around?").answer("The
Sun").createdAt(dateTimeFromString("2021-01-01T00:00:00Z")).build(),
new AiDialogueHistory.Builder().prompt("On Earth, where does the Sun
rise?").answer("east").createdAt(dateTimeFromString("2021-01-01T00:00:00Z")).build())).build())
- lang: node
label: Generate text
source: |-
await client.ai.createAiTextGen({
prompt: 'Paraphrase the documents',
items: [
new AiTextGenItemsField({
id: fileToAsk.id,
type: 'file' as AiTextGenItemsTypeField,
content:
'The Earth goes around the Sun. The Sun rises in the east in the morning.',
}),
],
dialogueHistory: [
{
prompt: 'What does the earth go around?',
answer: 'The Sun',
createdAt: dateTimeFromString('2021-01-01T00:00:00Z'),
} satisfies AiDialogueHistory,
{
prompt: 'On Earth, where does the Sun rise?',
answer: 'east',
createdAt: dateTimeFromString('2021-01-01T00:00:00Z'),
} satisfies AiDialogueHistory,
],
} satisfies AiTextGen);
- lang: python
label: Generate text
source: |-
client.ai.create_ai_text_gen(
"Paraphrase the documents",
[
CreateAiTextGenItems(
id=file_to_ask.id,
type=CreateAiTextGenItemsTypeField.FILE,
content="The Earth goes around the Sun. The Sun rises in the east in the morning.",
)
],
dialogue_history=[
AiDialogueHistory(
prompt="What does the earth go around?",
answer="The Sun",
created_at=date_time_from_string("2021-01-01T00:00:00Z"),
),
AiDialogueHistory(
prompt="On Earth, where does the Sun rise?",
answer="east",
created_at=date_time_from_string("2021-01-01T00:00:00Z"),
),
],
)
components:
schemas:
AiTextGen:
description: AI text gen request object.
type: object
properties:
prompt:
description: >-
The prompt provided by the client to be answered by the LLM. The
prompt's length is limited to 10000 characters.
type: string
example: Write an email to a client about the importance of public APIs.
items:
description: >-
The items to be processed by the LLM, often files. The array can
include **exactly one** element.
**Note**: Box AI handles documents with text representations up to
1MB in size. If the file size exceeds 1MB, the first 1MB of text
representation will be processed.
type: array
items:
required:
- id
- type
type: object
description: The item to be processed by the LLM.
properties:
id:
description: The ID of the item.
type: string
example: '123'
type:
description: The type of the item.
type: string
example: file
enum:
- file
content:
description: >-
The content to use as context for generating new text or
editing existing text.
example: This is file content that is relevant to the text gen request.
type: string
maxItems: 1
minItems: 1
uniqueItems: true
dialogue_history:
description: >-
The history of prompts and answers previously passed to the LLM.
This parameter provides the additional context to the LLM when
generating the response.
type: array
items:
$ref: '#/components/schemas/AiDialogueHistory'
ai_agent:
allOf:
- $ref: '#/components/schemas/AiTextGenAgent'
- description: The AI agent to be used for generating text.
required:
- prompt
- items
title: AI text gen request
x-box-tag: ai
AiResponse:
description: AI response.
type: object
properties:
answer:
description: The answer provided by the LLM.
type: string
example: Public APIs are important because of key and important reasons.
created_at:
description: >-
The ISO date formatted timestamp of when the answer to the prompt
was created.
type: string
format: date-time
example: '2012-12-12T10:53:43-08:00'
completion_reason:
description: The reason the response finishes.
type: string
example: done
ai_agent_info:
$ref: '#/components/schemas/AiAgentInfo'
required:
- answer
- created_at
title: AI response
x-box-resource-id: ai_response
x-box-tag: ai
x-box-variant: standard
x-box-variants:
- standard
- full
ClientError:
description: A generic error.
type: object
properties:
type:
description: The value will always be `error`.
type: string
example: error
enum:
- error
nullable: false
status:
description: The HTTP status of the response.
type: integer
format: int32
example: 400
nullable: false
code:
description: A Box-specific error code.
type: string
example: item_name_invalid
enum:
- created
- accepted
- no_content
- redirect
- not_modified
- bad_request
- unauthorized
- forbidden
- not_found
- method_not_allowed
- conflict
- precondition_failed
- too_many_requests
- internal_server_error
- unavailable
- item_name_invalid
- insufficient_scope
message:
description: A short message describing the error.
type: string
example: Method Not Allowed
nullable: false
context_info:
description: >-
A free-form object that contains additional context about the error.
The possible fields are defined on a per-endpoint basis. `message`
is only one example.
type: object
example:
message: Something went wrong
additionalProperties: {}
nullable: true
help_url:
description: A URL that links to more information about why this error occurred.
type: string
example: >-
https://developer.box.com/guides/api-calls/permissions-and-errors/common-errors/
nullable: false
request_id:
description: >-
A unique identifier for this response, which can be used when
contacting Box support.
type: string
example: abcdef123456
nullable: false
title: Client error
x-box-resource-id: client_error
AiDialogueHistory:
description: A context object that can hold prior prompts and answers.
type: object
properties:
prompt:
description: >-
The prompt previously provided by the client and answered by the
LLM.
type: string
example: Make my email about public APIs sound more professional.
answer:
description: The answer previously provided by the LLM.
type: string
example: >-
Here is the first draft of your professional email about public
APIs.
created_at:
description: >-
The ISO date formatted timestamp of when the previous answer to the
prompt was created.
type: string
format: date-time
example: '2012-12-12T10:53:43-08:00'
title: Dialogue history
AiTextGenAgent:
description: The AI agent to be used to handle the AI text generation request.
type: object
oneOf:
- $ref: '#/components/schemas/AiAgentReference'
- $ref: '#/components/schemas/AiAgentTextGen'
title: AI text gen request agent
AiAgentInfo:
description: The information on the models and processors used in the request.
type: object
properties:
models:
description: The models used for the request.
type: array
items:
type: object
properties:
name:
description: The name of the model used for the request.
type: string
example: azure__openai__text_embedding_ada_002
provider:
description: The provider that owns the model used for the request.
type: string
example: azure
supported_purpose:
description: >-
The supported purpose utilized by the model used for the
request.
type: string
example: embedding
processor:
description: The processor used for the request.
type: string
example: basic_text
title: The information on the models and processors used in the request.
AiAgentReference:
description: The AI agent used to handle queries.
type: object
properties:
type:
description: The type of AI agent used to handle queries.
type: string
example: ai_agent_id
enum:
- ai_agent_id
nullable: false
id:
description: >-
The ID of an Agent. This can be a numeric ID for custom agents (for
example, `14031`) or a unique identifier for pre-built agents (for
example, `enhanced_extract_agent` for the [Enhanced Extract
Agent](/guides/box-ai/ai-tutorials/extract-metadata-structured#enhanced-extract-agent)).
type: string
example: '14031'
nullable: false
required:
- type
title: AI agent reference
x-box-resource-id: ai_agent_reference
x-box-tag: ai
AiAgentTextGen:
description: The AI agent used for generating text.
type: object
properties:
type:
description: The type of AI agent used for generating text.
type: string
example: ai_agent_text_gen
enum:
- ai_agent_text_gen
nullable: false
basic_gen:
$ref: '#/components/schemas/AiAgentBasicGenTool'
required:
- type
title: AI agent for text generation requests
x-box-resource-id: ai_agent_text_gen
x-box-tag: ai
AiAgentBasicGenTool:
description: AI agent basic tool used to generate text.
type: object
allOf:
- $ref: '#/components/schemas/AiAgentLongTextToolTextGen'
- properties:
content_template:
description: >-
How the content should be included in a request to the LLM.
Input for `{content}` is optional, depending on the use.
type: string
example: '---{content}---'
title: AI agent basic text generation tool
x-box-tag: ai
AiAgentLongTextToolTextGen:
description: AI agent processor used to to handle longer text.
type: object
allOf:
- $ref: '#/components/schemas/AiAgentBasicTextToolTextGen'
- properties:
embeddings:
type: object
properties:
model:
description: The model used for the AI agent for calculating embeddings.
type: string
example: azure__openai__text_embedding_ada_002
strategy:
type: object
properties:
id:
description: >-
The strategy used for the AI agent for calculating
embeddings.
type: string
example: basic
num_tokens_per_chunk:
description: The number of tokens per chunk.
type: integer
example: 64
maximum: 512
minimum: 1
title: AI agent long text tool
x-box-tag: ai
AiAgentBasicTextToolTextGen:
description: AI agent processor used to handle basic text.
type: object
allOf:
- $ref: '#/components/schemas/AiAgentBasicTextToolBase'
- properties:
system_message:
description: >-
System messages aim at helping the LLM understand its role and
what it is supposed to do. The input for `{current_date}` is
optional, depending on the use.
type: string
example: You are a helpful travel assistant specialized in budget travel
prompt_template:
description: >-
The prompt template contains contextual information of the
request and the user prompt.
When using the `prompt_template` parameter, you **must include**
input for `{user_question}`. Inputs for `{current_date}` and
`{content}` are optional, depending on the use.
type: string
example: >-
It is `{current_date}`, and I have $8000 and want to spend a
week in the Azores. `{user_question}`
maxLength: 10000
pattern: \{user_question\}
title: AI agent basic text tool
x-box-tag: ai
AiAgentBasicTextToolBase:
description: AI agent processor used to handle basic text.
type: object
properties:
model:
description: >-
The model used for the AI agent for basic text. For specific model
values, see the [available models list](/guides/box-ai/ai-models).
type: string
example: azure__openai__gpt_4o_mini
num_tokens_for_completion:
description: The number of tokens for completion.
type: integer
example: 8400
minimum: 1
llm_endpoint_params:
$ref: '#/components/schemas/AiLlmEndpointParams'
title: AI agent basic text tool
x-box-tag: ai
AiLlmEndpointParams:
description: The parameters for the LLM endpoint specific to a model.
type: object
oneOf:
- $ref: '#/components/schemas/AiLlmEndpointParamsOpenAi'
- $ref: '#/components/schemas/AiLlmEndpointParamsGoogle'
- $ref: '#/components/schemas/AiLlmEndpointParamsAWS'
- $ref: '#/components/schemas/AiLlmEndpointParamsIBM'
title: AI LLM endpoint parameters
AiLlmEndpointParamsOpenAi:
description: AI LLM endpoint params OpenAI object.
type: object
properties:
type:
description: >-
The type of the AI LLM endpoint params object for OpenAI. This
parameter is **required**.
type: string
example: openai_params
enum:
- openai_params
nullable: false
temperature:
description: >-
What sampling temperature to use, between 0 and 2. Higher values
like 0.8 will make the output more random, while lower values like
0.2 will make it more focused and deterministic. We generally
recommend altering this or `top_p` but not both.
type: number
example: 0
maximum: 2
minimum: 0
nullable: true
top_p:
description: >-
An alternative to sampling with temperature, called nucleus
sampling, where the model considers the results of the tokens with
`top_p` probability mass. So 0.1 means only the tokens comprising
the top 10% probability mass are considered. We generally recommend
altering this or temperature but not both.
type: number
example: 1
maximum: 1
minimum: 0.1
nullable: true
frequency_penalty:
description: >-
A number between -2.0 and 2.0. Positive 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.
type: number
example: 1.5
maximum: 2
minimum: -2
nullable: true
presence_penalty:
description: >-
A number between -2.0 and 2.0. Positive values penalize new tokens
based on whether they appear in the text so far, increasing the
model's likelihood to talk about new topics.
type: number
example: 1.5
maximum: 2
minimum: -2
nullable: true
stop:
description: Up to 4 sequences where the API will stop generating further tokens.
type: string
example: <|im_end|>
nullable: true
required:
- type
title: AI LLM endpoint params OpenAI
x-box-resource-id: ai_llm_endpoint_params_openai
AiLlmEndpointParamsGoogle:
description: AI LLM endpoint params Google object.
type: object
properties:
type:
description: >-
The type of the AI LLM endpoint params object for Google. This
parameter is **required**.
type: string
example: google_params
enum:
- google_params
nullable: false
temperature:
description: >-
The temperature is used for sampling during response generation,
which occurs when `top-P` and `top-K` are applied. Temperature
controls the degree of randomness in the token selection.
type: number
example: 0
maximum: 2
minimum: 0
nullable: true
top_p:
description: >-
`Top-P` changes how the model selects tokens for output. Tokens are
selected from the most (see `top-K`) to least probable until the sum
of their probabilities equals the `top-P` value.
type: number
example: 1
maximum: 2
minimum: 0.1
nullable: true
top_k:
description: >-
`Top-K` changes how the model selects tokens for output. A low
`top-K` means the next selected token is the most probable among all
tokens in the model's vocabulary (also called greedy decoding),
while a high `top-K` means that the next token is selected from
among the three most probable tokens by using temperature.
type: number
example: 1
maximum: 2
minimum: 0.1
nullable: true
required:
- type
title: AI LLM endpoint params Google
x-box-resource-id: ai_llm_endpoint_params_google
AiLlmEndpointParamsAWS:
description: AI LLM endpoint params AWS object.
type: object
properties:
type:
description: >-
The type of the AI LLM endpoint params object for AWS. This
parameter is **required**.
type: string
example: aws_params
enum:
- aws_params
nullable: false
temperature:
description: >-
What sampling temperature to use, between 0 and 1. Higher values
like 0.8 will make the output more random, while lower values like
0.2 will make it more focused and deterministic. We generally
recommend altering this or `top_p` but not both.
type: number
example: 0.5
maximum: 1
minimum: 0
nullable: true
top_p:
description: >-
An alternative to sampling with temperature, called nucleus
sampling, where the model considers the results of the tokens with
`top_p` probability mass. So 0.1 means only the tokens comprising
the top 10% probability mass are considered. We generally recommend
altering this or temperature but not both.
type: number
example: 0.5
maximum: 1
minimum: 0
nullable: true
required:
- type
title: AI LLM endpoint params AWS
x-box-resource-id: ai_llm_endpoint_params_aws
AiLlmEndpointParamsIBM:
description: AI LLM endpoint params IBM object.
type: object
properties:
type:
description: >-
The type of the AI LLM endpoint params object for IBM. This
parameter is **required**.
type: string
example: ibm_params
enum:
- ibm_params
nullable: false
temperature:
description: >-
What sampling temperature to use, between 0 and 1. Higher values
like 0.8 will make the output more random, while lower values like
0.2 will make it more focused and deterministic. We generally
recommend altering this or `top_p` but not both.
type: number
example: 0.5
nullable: true
top_p:
description: >-
An alternative to sampling with temperature, called nucleus
sampling, where the model considers the results of the tokens with
`top_p` probability mass. So 0.1 means only the tokens comprising
the top 10% probability mass are considered. We generally recommend
altering this or temperature but not both.
type: number
example: 0.5
maximum: 1
minimum: 0.1
nullable: true
top_k:
description: >-
`Top-K` changes how the model selects tokens for output. A low
`top-K` means the next selected token is the most probable among all
tokens in the model's vocabulary (also called greedy decoding),
while a high `top-K` means that the next token is selected from
among the three most probable tokens by using temperature.
type: number
example: 1
nullable: true
required:
- type
title: AI LLM endpoint params IBM
x-box-resource-id: ai_llm_endpoint_params_ibm
securitySchemes:
OAuth2Security:
type: oauth2
flows:
authorizationCode:
authorizationUrl: https://account.box.com/api/oauth2/authorize
tokenUrl: https://api.box.com/oauth2/token
scopes:
root_readonly: Read all files and folders stored in Box
root_readwrite: Read and write all files and folders stored in Box
manage_app_users: Provision and manage app users
manage_managed_users: Provision and manage managed users
manage_groups: Manage an enterprise's groups
manage_webhook: Create webhooks programmatically through the API
manage_enterprise_properties: Manage enterprise properties
manage_data_retention: Manage data retention polices
manage_legal_hold: Manage Legal Holds
````