ASAPP AutoCompose API
Generate next-message suggestions for agents, spelling correction, profanity checks, and custom responses.
OpenAPI Specification
openapi: 3.0.1
info:
title: AutoCompose API
description: >
Autocompose API to suggest the next agent message.
Suggestions are based on the conversation history, conversation metadata,
and
the in-progress message text the agent has already typed into the composer.
version: 0.0.3
servers:
- url: https://api.sandbox.asapp.com
tags:
- name: AutoCompose
description: Improve agent productivity with AutoCompose API
security:
- API-ID: []
API-Secret: []
paths:
/autocompose/v1/conversations/{conversationId}/suggestions:
parameters:
- name: conversationId
description: The identifier for a conversation.
in: path
required: true
schema:
type: string
pattern: ^[A-Z0-9]+$
post:
tags:
- AutoCompose
operationId: getSuggestions
summary: Generate suggestions
description: >
Get suggestions for the next agent message in the conversation.
There are several times when this should be called:
- when an agent joins the conversation,
- after a message is sent by either the customer or the agent,
- and as the agent is typing in the composer (to enable completing the
agent's in-progress message).
Optionally, add a message to the conversation.
requestBody:
description: The parameters for getting suggestions for the next agent message.
content:
application/json:
schema:
description: >-
A suggestions request with an optional message to add to the
conversation
type: object
properties:
query:
description: text the agent has already entered into the composer
type: string
message:
description: Represents a single message within a conversation.
type: object
properties:
text:
type: string
minLength: 1
description: The content of the message.
sender:
description: Information about the participant who sent the message.
type: object
properties:
role:
description: The role of the participant in the conversation.
type: string
enum:
- agent
- customer
- system
externalId:
type: string
description: >-
The unique identifier for the participant in your
external system. This should be consistent across
all interactions for the same individual.
required:
- role
- externalId
timestamp:
type: string
format: date-time
description: >-
The time when the message was sent. Include the
timezone, otherwise UTC will be assumed.
required:
- text
- sender
- timestamp
example:
text: Hello, I would like to upgrade my internet plan to GOLD.
sender:
role: agent
externalId: 123
timestamp: '2021-11-23T12:13:14.555Z'
example:
query: Hello, how can
message:
text: Hello, I would like to upgrade my internet plan to GOLD.
sender:
role: customer
externalId: customer-x
timestamp: '2023-06-01T19:16:55.706Z'
responses:
'200':
description: Successfully fetched suggestions for the conversation
content:
application/json:
schema:
description: Suggestions for next agent message
type: object
properties:
id:
description: >-
ID for the suggestions that were returned (to be used in
analytics events)
type: string
message:
description: Response for messages
type: object
properties:
id:
type: string
description: Message ID to be used for analytics
example:
id: 01BX5ZZKBKACTAV9WEVGEMMVS1
suggestions:
type: array
items:
description: Suggestion
type: object
properties:
text:
type: string
description: Suggestion text for the agent
example: Hello John, how can I help you?
templateText:
type: string
description: >-
Suggestion text with placeholders for metadata
instead of the values filled in
example: Hello {NAME}, how can I help you?
title:
type: string
description: The title of the response
example: Greeting
phraseCompletion:
description: >-
Completion of the phrase the agent is currently typing, to
be displayed inline in the composer
type: object
properties:
text:
type: string
description: >-
Phrase completion text for the agent (this may be
empty if the model does not have a confident
prediction for how to complete the message)
example: Hello, how can I help you?
autopilotMessage:
description: Recommended message for the agent to autopilot send.
type: object
properties:
text:
type: string
description: Text of the message
example: Hello John, how can I help you?
templateText:
type: string
description: >-
Text of the message with placeholders for metadata
instead of the values filled in
example: Hello {NAME}, how can I help you?
title:
type: string
description: The title of the message
example: Greeting
delaySeconds:
type: integer
description: >-
Delay before the message is sent to the customer,
during which a preview of the message is shown to the
agent and they can choose to cancel it
example: 5
triggerAfterSeconds:
type: integer
description: >-
Delay before showing the preview and starting the
countdown to send the message
example: 0
'400':
description: 400 - Bad request
content:
application/json:
schema:
description: Bad request response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 400-01
message: Bad request
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'401':
description: 401 - Unauthorized
content:
application/json:
schema:
description: Unauthorized response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 401-01
message: Unauthorized
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'403':
description: 403 - Forbidden
content:
application/json:
schema:
description: Forbidden response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 403-01
message: Forbidden Response
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'404':
description: 404 - Not Found
content:
application/json:
schema:
description: Not Found response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 404-01
message: Not Found
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'409':
description: 409 - Conflict
content:
application/json:
schema:
description: Conflict response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 409-01
message: Conflict
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'413':
description: 413 - Request Entity Too Large
content:
application/json:
schema:
description: Request Entity Too Large response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 413-01
message: Request Entity Too Large
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'422':
description: 422 - Unprocessable Entity
content:
application/json:
schema:
description: Unprocessable Entity response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 422-01
message: Unprocessable Entity
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'429':
description: 429 - Too Many Requests
content:
application/json:
schema:
description: Too Many Requests response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 429-01
message: Too Many Requests
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'503':
description: 503 - Service Unavailable
content:
application/json:
schema:
description: Service Unavailable response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 503-01
message: Service Unavailable
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
default:
description: 500 - Internal Server Error
content:
application/json:
schema:
description: Default error response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 500-01
message: Internal server error
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
/autocompose/v1/profanity/evaluation:
post:
tags:
- AutoCompose
operationId: getEvaluation
summary: Evaluate profanity
description: >
Get an evaluation of a text to verify if it contains profanity,
obscenity or other unwanted words. This service should be called before
sending a message to prevent the agent from sending profanities in the
chat.
requestBody:
description: >-
The parameters for getting a evaluation result of the text typed by
the user.
content:
application/json:
schema:
type: object
properties:
text:
description: text to be checked for profanity
type: string
example:
text: This is a perfectly decent sentence.
responses:
'200':
description: Successfully fetched a evaluation result of the sentence.
content:
application/json:
schema:
description: ProfanityResult
type: object
properties:
hasProfanity:
description: >-
A boolean value describing if the text contains a curse
word.
type: boolean
example: false
'400':
description: 400 - Bad request
content:
application/json:
schema:
description: Bad request response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 400-01
message: Bad request
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'401':
description: 401 - Unauthorized
content:
application/json:
schema:
description: Unauthorized response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 401-01
message: Unauthorized
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'403':
description: 403 - Forbidden
content:
application/json:
schema:
description: Forbidden response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 403-01
message: Forbidden Response
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'404':
description: 404 - Not Found
content:
application/json:
schema:
description: Not Found response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 404-01
message: Not Found
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'409':
description: 409 - Conflict
content:
application/json:
schema:
description: Conflict response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 409-01
message: Conflict
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'413':
description: 413 - Request Entity Too Large
content:
application/json:
schema:
description: Request Entity Too Large response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 413-01
message: Request Entity Too Large
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'422':
description: 422 - Unprocessable Entity
content:
application/json:
schema:
description: Unprocessable Entity response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 422-01
message: Unprocessable Entity
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'429':
description: 429 - Too Many Requests
content:
application/json:
schema:
description: Too Many Requests response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 429-01
message: Too Many Requests
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
'503':
description: 503 - Service Unavailable
content:
application/json:
schema:
description: Service Unavailable response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 503-01
message: Service Unavailable
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
default:
description: 500 - Internal Server Error
content:
application/json:
schema:
description: Default error response
type: object
properties:
error:
example:
requestId: 8e033668-9f1a-11ec-b909-0242ac120002
code: 500-01
message: Internal server error
description: Error details
type: object
properties:
requestId:
type: string
description: Unique ID of the failing request
message:
type: string
description: Error message
code:
type: string
description: Error code
required:
- requestId
- message
/autocompose/v1/spellcheck/correction:
post:
tags:
- AutoCompose
operationId: getSpellingCorrection
summary: Check for spelling mistakes
description: >
Get spelling correction for a message as it is being typed, if there is
a misspelling. Only the current word
will be corrected, once it's fully typed (so it is recommended to call
this endpoint after space characters).
requestBody:
description: >-
The parameters for getting a spelling correction as a message is being
typed.
required: true
content:
application/json:
schema:
type: object
properties:
text:
description: >-
Text being typed by the agent, already entered into the
composer
type: string
typingEvent:
type: object
properties:
cursorStart:
description: >-
Position of the cursor before the typing event, to
identify if a word is being typed
type: integer
cursorEnd:
description: >-
Position of the cursor after the typing event, to
identify if a word is being typed
type: integer
required:
- cursorStart
- cursorEnd
userDictionary:
description: >-
Words that should not be corrected if they are present on
the text
type: array
items:
type: string
language:
description: >-
Optional IETF language tag of the text. If not provided, the
default language will be assumed which most likely will be
English but it could be a different language depending on
the customer setup.
type: string
required:
- text
- typingEvent
example:
text: 'How is tihs '
typingEvent:
cursorStart: 11
cursorEnd: 12
userDictionary:
- Hellooo
language: en
responses:
'200':
description: Successfully checked for a spelling mistake.
content:
application/json:
schema:
type: object
propert
# --- truncated at 32 KB (285 KB total) ---
# Full source: https://raw.githubusercontent.com/api-evangelist/asapp/refs/heads/main/openapi/asapp-autocompose-openapi.yml