> ## Documentation Index
> Fetch the complete documentation index at: https://docs.reply.io/llms.txt
> Use this file to discover all available pages before exploring further.
# Create an AI learning
>
**Coming soon.** This endpoint will be available by early October 2026.
_Requires the `ai-sdr:write` scope (or a broader one that includes it)._
Use this endpoint when you need a new standing rule for how the AI writes. text is required and
limited to 1000 characters, and categories says where the rule applies — personalizedMessage for
outbound drafting, replyHandling for answering incoming mail, or both. scope decides who the rule
binds: user for yourself across all your sequences, sequence for named sequences only,
workspace for everyone in your team, or organization for every workspace in the org.
sequenceIds names the sequences — at most 50 and no repeats, each one you can see — and is
required for the sequence scope and rejected for every other one. Creating a workspace or
organization learning needs the matching permission, and an organization learning needs you to
belong to one — without either the call is refused and nothing is written. A learning created
here is always enabled and counts as manually authored, so it can be refused with a conflict when
those categories and scope already hold as many manually authored learnings as they may; the
message says which budget is full. The created learning is returned in full with its assigned
id — unless the service cannot be reached to read it straight back, in which case the response
carries the assigned id and the fields you sent rather than the stored learning. Contradictions
with your other learnings are looked for after the learning is stored, so conflictState comes
back as none here — read the learning again a moment later to see its conflict standing.
The AI Learnings feature must be enabled for your team.
## OpenAPI
````yaml /api-reference/bundled.yaml post /v3/ai-sdr/learnings
openapi: 3.1.0
info:
version: 3.0.0
title: Reply API
description: API for managing email sequences, contacts, and automation workflows
servers:
- url: https://api.reply.io
security:
- bearerAuth: []
tags:
- name: Contacts
description: Manage individual contacts
- name: Contact Lists
description: Manage contact lists and contact membership
- name: Accounts
description: Manage accounts (companies/organizations)
- name: Account Lists
description: Manage account lists and account membership
- name: Custom Fields
description: Manage custom contact fields
- name: Prompt Actions
description: Save reusable AI prompts and run them on contacts to fill custom fields
- name: Contact Blacklist Rules
description: Manage blacklist rules for domains, emails, and email exceptions
- name: Live Data
description: >-
Find new contacts via Live Data searches and read typeahead values for the
Live Data / Autopilot filter sidebar
- name: Contact Enrichment
description: >-
Enrich contacts with emails, phone numbers, LinkedIn data, and AI-filled
custom fields
- name: Email Validations
description: Estimate and schedule asynchronous email validation jobs
- name: AI SDR Web Search
description: Find contacts via AI-driven web search and review past searches
- name: Sequences
description: Manage email automation sequences
- name: Sequence Steps
description: Manage individual steps in sequences
- name: Sequence Contacts
description: Manage contacts within sequences
- name: Sequence Folders
description: Organize sequences into folders
- name: Email Accounts
description: Manage email accounts used for sending and receiving
- name: Sequence Email Accounts
description: Manage email accounts linked to sequences
- name: LinkedIn Accounts
description: Manage LinkedIn accounts for outreach
- name: Sequence LinkedIn Accounts
description: Manage LinkedIn accounts linked to sequences
- name: Sequence Templates
description: Manage sequence templates
- name: Email Templates
description: Manage email templates and template folders
- name: Voices
description: Manage cloned voices used for AI voice message steps
- name: Schedules
description: Manage email send schedules and timing
- name: Holiday Calendars
description: Manage holiday calendars for scheduling
- name: Inbox
description: >-
Manage inbox threads — list/filter, read state, replies, category
assignment, and meeting-intent flagging
- name: Inbox Categories
description: Manage per-team inbox thread categories and assign threads to them
- name: Direct Outreach
description: >-
Send one-off outreach directly to a contact outside of any sequence —
direct emails and LinkedIn connection requests, messages, InMails, and
voice messages
- name: Tasks
description: Manage tasks and to-do items
- name: Reports
description: >-
Generate and access performance reports across email, calls, tasks,
LinkedIn, and team performance
- name: AI SDR Sequences
description: >-
Manage AI SDR sequences and their AI SDR-specific settings — create, read
settings, partial update, autopilot enable/disable/force-start, approval
mode, generated step types, and playbook/knowledge-base connections
- name: AI SDR Strategist
description: Trigger AI Strategist runs
- name: AI SDR Playbooks
description: >-
Manage AI SDR playbooks — tone, voice, and style guides applied during
personalized message generation
- name: AI SDR Knowledge Bases
description: >-
Manage AI SDR knowledge bases — collections of documents, links, reply
handlers, and reengagement cards that inform the agent's responses
- name: AI SDR Offers
description: >-
Manage AI SDR offers — bundles of company-context inputs (ICP, pain
points, value propositions, etc.) used to personalize outreach
- name: AI SDR Pending Approvals
description: >-
Review, send, regenerate, and provide feedback on AI-generated messages
awaiting human approval
- name: AI SDR Sequence Preview
description: >-
Read and regenerate per-contact previews of the messages a sequence will
send, and provide feedback on preview messages
- name: AI SDR Insights
description: Read AI SDR insights for sequence contacts
- name: AI SDR Intent Signals
description: >-
Read Reply industry IDs and technology slugs used in AI SDR intent-signal
configuration (typeahead)
- name: AI Prompts
description: >-
Manage the AI prompts used to configure sequence steps, and preview the
output a prompt produces
- name: AI SDR Learnings
description: >-
Manage the standing rules that shape how the AI writes — list, read,
create, partially update and delete AI learnings
- name: Registration
description: Register the contact that domain purchases are filed under
- name: Domains
description: Find domains available to buy for outreach
- name: User Account
description: Account information and authentication verification
- name: Settings
description: Manage team and user settings
- name: Billing
description: Subscription details and monthly active contacts usage
- name: Webhooks
description: Manage webhook subscriptions and inspect delivery history
- name: Background Jobs
description: >-
Track and cancel asynchronous background operations (e.g., email
validation)
- name: Attachments
description: >-
Upload file attachments used across email templates, sequence steps, and
direct emails
paths:
/v3/ai-sdr/learnings:
post:
tags:
- AI SDR Learnings
summary: Create an AI learning
description: >-
**Coming soon.** This endpoint will be available by early October 2026.
_Requires the `ai-sdr:write` scope (or a broader one that
includes it)._
Use this endpoint when you need a new standing rule for how the AI
writes. text is required and
limited to 1000 characters, and categories says where the rule applies —
personalizedMessage for
outbound drafting, replyHandling for answering incoming mail, or both.
scope decides who the rule
binds: user for yourself across all your sequences, sequence for named
sequences only,
workspace for everyone in your team, or organization for every workspace
in the org.
sequenceIds names the sequences — at most 50 and no repeats, each one
you can see — and is
required for the sequence scope and rejected for every other one.
Creating a workspace or
organization learning needs the matching permission, and an organization
learning needs you to
belong to one — without either the call is refused and nothing is
written. A learning created
here is always enabled and counts as manually authored, so it can be
refused with a conflict when
those categories and scope already hold as many manually authored
learnings as they may; the
message says which budget is full. The created learning is returned in
full with its assigned
id — unless the service cannot be reached to read it straight back, in
which case the response
carries the assigned id and the fields you sent rather than the stored
learning. Contradictions
with your other learnings are looked for after the learning is stored,
so conflictState comes
back as none here — read the learning again a moment later to see its
conflict standing.
The AI Learnings feature must be enabled for your team.
operationId: CreateAiLearning
requestBody:
required: true
content:
application/json:
schema:
type: object
title: AiLearningCreationRequest
description: >-
Request body for creating an AI learning. A learning created
through the API is always enabled
and counts as manually authored.
properties:
text:
type: string
minLength: 1
maxLength: 1000
description: The standing rule the AI should follow
categories:
type: array
minItems: 1
uniqueItems: true
description: >-
Where the rule applies. Send one or both values, without
repeats.
items:
type: string
title: AI Learning Category
description: >-
Where a learning applies. `personalizedMessage` shapes
outbound message drafting, `replyHandling`
shapes answers to incoming mail. A learning can carry
both.
enum:
- personalizedMessage
- replyHandling
scope:
description: >-
Who the learning binds. `workspace` and `organization` need
the matching permission, and
`organization` also needs you to belong to an organization.
type: string
title: AI Learning Scope
enum:
- organization
- workspace
- user
- sequence
sequenceIds:
type: array
maxItems: 50
uniqueItems: true
description: >-
Sequences the learning is limited to, each one a sequence
you can see. Required when `scope`
is `sequence` and rejected for every other scope.
items:
type: integer
minimum: 1
required:
- text
- categories
- scope
example:
text: Keep subject lines under six words.
categories:
- personalizedMessage
scope: sequence
sequenceIds:
- 42
- 57
responses:
'201':
description: The learning was created
content:
application/json:
schema:
type: object
title: AiLearning
description: >-
A standing rule that shapes how the AI writes, together with
the learnings it contradicts.
properties:
id:
type: integer
description: Unique identifier of the learning
readOnly: true
text:
type: string
description: The standing rule the AI follows
readOnly: true
categories:
type: array
description: Where the rule applies
items:
type: string
title: AI Learning Category
description: >-
Where a learning applies. `personalizedMessage` shapes
outbound message drafting, `replyHandling`
shapes answers to incoming mail. A learning can carry
both.
enum:
- personalizedMessage
- replyHandling
readOnly: true
source:
description: Whether the AI derived the learning or a person wrote it
type: string
title: AI Learning Source
enum:
- aiGenerated
- manual
scope:
description: Who the learning binds
type: string
title: AI Learning Scope
enum:
- organization
- workspace
- user
- sequence
sequenceIds:
type: array
description: >-
Sequences the learning is limited to. Empty unless `scope`
is `sequence`.
items:
type: integer
readOnly: true
isEnabled:
type: boolean
description: Whether the learning is switched on
readOnly: true
createdAt:
type: string
format: date-time
description: When the learning was created (UTC)
readOnly: true
updatedAt:
type: string
format: date-time
nullable: true
description: >-
When the learning was last changed (UTC). `null` if it has
never been changed.
readOnly: true
ownerUserId:
type: integer
nullable: true
description: >-
User ID of the learning's owner. `null` when the owner is
no longer a Reply user.
readOnly: true
isOwn:
type: boolean
description: >-
Whether the learning is yours. `false` for a learning a
team or organization admin set for everyone.
readOnly: true
conflictState:
description: >-
How the learning stands against the learnings it
contradicts. An `unresolved` learning is not applied.
type: string
title: AI Learning Conflict State
enum:
- none
- overridden
- takesPrecedence
- unresolved
conflicts:
type: array
description: >-
Every other learning this one contradicts. Empty when
`conflictState` is `none`.
items:
type: object
title: AiLearningConflict
description: >-
Another AI learning that contradicts the one being read,
and which of the two currently applies.
properties:
otherLearningId:
type: integer
description: ID of the contradicting learning
readOnly: true
otherText:
type: string
description: Text of the contradicting learning
readOnly: true
categories:
type: array
description: Categories the contradicting learning applies to
items:
type: string
title: AI Learning Category
description: >-
Where a learning applies. `personalizedMessage`
shapes outbound message drafting, `replyHandling`
shapes answers to incoming mail. A learning can
carry both.
enum:
- personalizedMessage
- replyHandling
readOnly: true
state:
description: Which of the two learnings currently applies
type: string
title: AI Learning Conflict State
enum:
- none
- overridden
- takesPrecedence
- unresolved
reason:
type: string
description: Why the two learnings contradict each other
readOnly: true
example:
otherLearningId: 5512
otherText: Always open with a short greeting.
categories:
- personalizedMessage
state: overridden
reason: One learning forbids a greeting the other requires.
readOnly: true
example:
id: 5498
text: Never open a message with "Hello".
categories:
- personalizedMessage
source: manual
scope: user
sequenceIds: []
isEnabled: true
createdAt: '2026-09-02T09:14:07Z'
updatedAt: '2026-09-11T16:41:55Z'
ownerUserId: 1234
isOwn: true
conflictState: takesPrecedence
conflicts:
- otherLearningId: 5512
otherText: Always open with a short greeting.
categories:
- personalizedMessage
state: takesPrecedence
reason: One learning forbids a greeting the other requires.
'400':
description: >-
The request body failed validation, or a sequence in sequenceIds is
not one you can see.
content:
application/problem+json:
schema:
oneOf:
- allOf:
- allOf:
- type: object
title: Problem Details
description: >-
Bare RFC 9457 problem-details envelope. Returned
by middleware-level errors
that don't carry domain context: 401 Unauthorized
(auth middleware),
429 Too Many Requests (rate-limit middleware), and
route-level 404 / 405 /
415 (framework middleware).
Business and validation responses extend this
envelope and add additional
fields — see `business-problem.model.yaml` (adds
`code` slug) and
`validation-problem.model.yaml` (adds `errors[]`
array).
properties:
title:
type: string
description: Short, human-readable summary of the problem.
status:
type: integer
description: HTTP status code.
minimum: 100
maximum: 599
detail:
type: string
description: >-
Human-readable explanation specific to this
occurrence.
- type: object
properties:
errors:
type: array
description: >-
List of field-level validation errors. Always
non-empty when this
envelope is returned. Each entry pins a single
offending field
via JSON Pointer plus a sanitized detail
string.
items:
type: object
title: Validation Error
description: A single field-level validation error.
properties:
pointer:
type: string
description: >-
JSON Pointer (RFC 6901) to the offending
field — e.g.
`/steps/0/subject`.
* An empty string (`""`) means the error
applies to the whole
request body (e.g. body is missing or unparseable).
* For route or query parameter failures
the pointer is the parameter
name (e.g. `id`, `top`).
example: /steps/0/subject
detail:
type: string
description: >-
Sanitized, human-readable explanation of
this field's error. One of a
small set of templates — `"Field is
required."`, `"Value has an
invalid type."`, `"Value has an invalid
format."`, `"Request body is
not valid JSON."`, `"The request body is
required and cannot be
empty."` — or a FluentValidator message
on body endpoints.
example: Field is required.
title: Validation Problem
description: >-
Input-validation error response at 400. Returned when
the request body
fails binding, FluentValidator rules, or when
route/query parameter
attribute validation (`[Range]`, `[Required]`) fails.
Route, query, and
body errors are combined into a single `errors[]`
array — clients should
not assume one error per request.
- example:
title: Validation failed
status: 400
detail: The request body contains validation errors.
errors:
- pointer: /name
detail: Field is required.
- allOf:
- allOf:
- type: object
title: Problem Details
description: >-
Bare RFC 9457 problem-details envelope. Returned
by middleware-level errors
that don't carry domain context: 401 Unauthorized
(auth middleware),
429 Too Many Requests (rate-limit middleware), and
route-level 404 / 405 /
415 (framework middleware).
Business and validation responses extend this
envelope and add additional
fields — see `business-problem.model.yaml` (adds
`code` slug) and
`validation-problem.model.yaml` (adds `errors[]`
array).
properties:
title:
type: string
description: Short, human-readable summary of the problem.
status:
type: integer
description: HTTP status code.
minimum: 100
maximum: 599
detail:
type: string
description: >-
Human-readable explanation specific to this
occurrence.
- type: object
properties:
code:
type: string
description: >-
Stable, machine-readable error slug in the
form
`"."`.
* `resource` is the camelCased domain — e.g.
`sequence`,
`contact`, `inboxThread`, `blacklistDomainRule`.
* `variant` is the camelCased specific failure
mode — e.g.
`notFound`, `forbidden`, `duplicateName`, `globalRuleReadOnly`.
Use `code` for programmatic error handling;
use `detail` for
user-facing messages. Slugs are stable across
server-side enum
reorderings and never change meaning under a
given resource.
pattern: ^[a-z][a-zA-Z0-9]*(\.[a-z][a-zA-Z0-9]*)+$
example: sequence.notFound
title: Business Problem
description: >-
Domain error response carrying a stable, namespaced
error slug.
Emitted for most 4xx responses (business 400, 403,
404, 409, …) and
for 503 when a downstream dependency is unavailable.
- example:
title: Bad Request
status: 400
detail: Pagination parameters are invalid.
code: sequence.invalidPagination
examples:
text_too_long:
summary: text is longer than 1000 characters
value:
title: Validation failed
status: 400
detail: The request body contains validation errors.
errors:
- pointer: /text
detail: text must not exceed 1000 characters.
sequence_ids_required:
summary: scope is sequence but sequenceIds is missing
value:
title: Validation failed
status: 400
detail: The request body contains validation errors.
errors:
- pointer: /sequenceIds
detail: sequenceIds is required when scope is 'sequence'.
sequences_not_found:
summary: A sequence in sequenceIds is not one you can see
value:
title: Bad Request
status: 400
detail: Sequence(s) 42, 57 not found.
code: aiLearning.invalidParameter
'401':
description: >-
Unauthorized. The response body is empty; check the
`WWW-Authenticate` header for the expected scheme.
content:
application/problem+json:
schema:
allOf:
- type: object
title: Problem Details
description: >-
Bare RFC 9457 problem-details envelope. Returned by
middleware-level errors
that don't carry domain context: 401 Unauthorized (auth
middleware),
429 Too Many Requests (rate-limit middleware), and
route-level 404 / 405 /
415 (framework middleware).
Business and validation responses extend this envelope and
add additional
fields — see `business-problem.model.yaml` (adds `code`
slug) and
`validation-problem.model.yaml` (adds `errors[]` array).
properties:
title:
type: string
description: Short, human-readable summary of the problem.
status:
type: integer
description: HTTP status code.
minimum: 100
maximum: 599
detail:
type: string
description: >-
Human-readable explanation specific to this
occurrence.
- example:
title: Unauthorized
status: 401
detail: Authentication credentials are missing or invalid.
'403':
description: >-
The AI Learnings feature is not enabled for your team, or you may
not create a learning with the requested scope.
content:
application/problem+json:
schema:
allOf:
- allOf:
- type: object
title: Problem Details
description: >-
Bare RFC 9457 problem-details envelope. Returned by
middleware-level errors
that don't carry domain context: 401 Unauthorized
(auth middleware),
429 Too Many Requests (rate-limit middleware), and
route-level 404 / 405 /
415 (framework middleware).
Business and validation responses extend this envelope
and add additional
fields — see `business-problem.model.yaml` (adds
`code` slug) and
`validation-problem.model.yaml` (adds `errors[]`
array).
properties:
title:
type: string
description: Short, human-readable summary of the problem.
status:
type: integer
description: HTTP status code.
minimum: 100
maximum: 599
detail:
type: string
description: >-
Human-readable explanation specific to this
occurrence.
- type: object
properties:
code:
type: string
description: >-
Stable, machine-readable error slug in the form
`"."`.
* `resource` is the camelCased domain — e.g.
`sequence`,
`contact`, `inboxThread`, `blacklistDomainRule`.
* `variant` is the camelCased specific failure
mode — e.g.
`notFound`, `forbidden`, `duplicateName`, `globalRuleReadOnly`.
Use `code` for programmatic error handling; use
`detail` for
user-facing messages. Slugs are stable across
server-side enum
reorderings and never change meaning under a given
resource.
pattern: ^[a-z][a-zA-Z0-9]*(\.[a-z][a-zA-Z0-9]*)+$
example: sequence.notFound
title: Business Problem
description: >-
Domain error response carrying a stable, namespaced error
slug.
Emitted for most 4xx responses (business 400, 403, 404,
409, …) and
for 503 when a downstream dependency is unavailable.
- example:
title: Forbidden
status: 403
detail: You do not have permission to access this resource.
code: sequence.forbidden
examples:
feature_unavailable:
summary: The feature is not enabled for your team
value:
title: Forbidden
status: 403
detail: AI Learnings is not available for your team
code: aiLearning.forbidden
scopes_unavailable:
summary: >-
Your team's plan does not include team or organization
learnings
value:
title: Forbidden
status: 403
detail: >-
Sharing AI Learnings beyond your own account is not
available for your team
code: aiLearning.forbidden
scope_not_permitted:
summary: >-
You lack the permission for the workspace or organization
scope
value:
title: Forbidden
status: 403
detail: >-
You do not have permission to share AI Learnings with your
team or organization
code: aiLearning.forbidden
'404':
description: AI learning not found
content:
application/problem+json:
schema:
allOf:
- allOf:
- type: object
title: Problem Details
description: >-
Bare RFC 9457 problem-details envelope. Returned by
middleware-level errors
that don't carry domain context: 401 Unauthorized
(auth middleware),
429 Too Many Requests (rate-limit middleware), and
route-level 404 / 405 /
415 (framework middleware).
Business and validation responses extend this envelope
and add additional
fields — see `business-problem.model.yaml` (adds
`code` slug) and
`validation-problem.model.yaml` (adds `errors[]`
array).
properties:
title:
type: string
description: Short, human-readable summary of the problem.
status:
type: integer
description: HTTP status code.
minimum: 100
maximum: 599
detail:
type: string
description: >-
Human-readable explanation specific to this
occurrence.
- type: object
properties:
code:
type: string
description: >-
Stable, machine-readable error slug in the form
`"."`.
* `resource` is the camelCased domain — e.g.
`sequence`,
`contact`, `inboxThread`, `blacklistDomainRule`.
* `variant` is the camelCased specific failure
mode — e.g.
`notFound`, `forbidden`, `duplicateName`, `globalRuleReadOnly`.
Use `code` for programmatic error handling; use
`detail` for
user-facing messages. Slugs are stable across
server-side enum
reorderings and never change meaning under a given
resource.
pattern: ^[a-z][a-zA-Z0-9]*(\.[a-z][a-zA-Z0-9]*)+$
example: sequence.notFound
title: Business Problem
description: >-
Domain error response carrying a stable, namespaced error
slug.
Emitted for most 4xx responses (business 400, 403, 404,
409, …) and
for 503 when a downstream dependency is unavailable.
- example:
title: Not Found
status: 404
detail: The requested resource was not found.
code: sequence.notFound
example:
title: Not Found
status: 404
detail: AI learning not found
code: aiLearning.notFound
'409':
description: >-
Those categories and scope already hold as many manually authored
learnings as they may. The detail says which limit is reached.
content:
application/problem+json:
schema:
allOf:
- allOf:
- type: object
title: Problem Details
description: >-
Bare RFC 9457 problem-details envelope. Returned by
middleware-level errors
that don't carry domain context: 401 Unauthorized
(auth middleware),
429 Too Many Requests (rate-limit middleware), and
route-level 404 / 405 /
415 (framework middleware).
Business and validation responses extend this envelope
and add additional
fields — see `business-problem.model.yaml` (adds
`code` slug) and
`validation-problem.model.yaml` (adds `errors[]`
array).
properties:
title:
type: string
description: Short, human-readable summary of the problem.
status:
type: integer
description: HTTP status code.
minimum: 100
maximum: 599
detail:
type: string
description: >-
Human-readable explanation specific to this
occurrence.
- type: object
properties:
code:
type: string
description: >-
Stable, machine-readable error slug in the form
`"."`.
* `resource` is the camelCased domain — e.g.
`sequence`,
`contact`, `inboxThread`, `blacklistDomainRule`.
* `variant` is the camelCased specific failure
mode — e.g.
`notFound`, `forbidden`, `duplicateName`, `globalRuleReadOnly`.
Use `code` for programmatic error handling; use
`detail` for
user-facing messages. Slugs are stable across
server-side enum
reorderings and never change meaning under a given
resource.
pattern: ^[a-z][a-zA-Z0-9]*(\.[a-z][a-zA-Z0-9]*)+$
example: sequence.notFound
title: Business Problem
description: >-
Domain error response carrying a stable, namespaced error
slug.
Emitted for most 4xx responses (business 400, 403, 404,
409, …) and
for 503 when a downstream dependency is unavailable.
- example:
title: Conflict
status: 409
detail: A resource with this name already exists.
code: sequence.duplicateName
example:
title: Conflict
status: 409
detail: >-
The limit of manually authored learnings for this scope is
reached.
code: aiLearning.limitReached
'502':
description: The AI Learnings service is temporarily unavailable.
content:
application/problem+json:
schema:
allOf:
- allOf:
- type: object
title: Problem Details
description: >-
Bare RFC 9457 problem-details envelope. Returned by
middleware-level errors
that don't carry domain context: 401 Unauthorized
(auth middleware),
429 Too Many Requests (rate-limit middleware), and
route-level 404 / 405 /
415 (framework middleware).
Business and validation responses extend this envelope
and add additional
fields — see `business-problem.model.yaml` (adds
`code` slug) and
`validation-problem.model.yaml` (adds `errors[]`
array).
properties:
title:
type: string
description: Short, human-readable summary of the problem.
status:
type: integer
description: HTTP status code.
minimum: 100
maximum: 599
detail:
type: string
description: >-
Human-readable explanation specific to this
occurrence.
- type: object
properties:
code:
type: string
description: >-
Stable, machine-readable error slug in the form
`"."`.
* `resource` is the camelCased domain — e.g.
`sequence`,
`contact`, `inboxThread`, `blacklistDomainRule`.
* `variant` is the camelCased specific failure
mode — e.g.
`notFound`, `forbidden`, `duplicateName`, `globalRuleReadOnly`.
Use `code` for programmatic error handling; use
`detail` for
user-facing messages. Slugs are stable across
server-side enum
reorderings and never change meaning under a given
resource.
pattern: ^[a-z][a-zA-Z0-9]*(\.[a-z][a-zA-Z0-9]*)+$
example: sequence.notFound
title: Business Problem
description: >-
Domain error response carrying a stable, namespaced error
slug.
Emitted for most 4xx responses (business 400, 403, 404,
409, …) and
for 503 when a downstream dependency is unavailable.
- example:
title: Bad Request
status: 400
detail: Pagination parameters are invalid.
code: sequence.invalidPagination
example:
title: Bad Gateway
status: 502
detail: The AI Learnings service is temporarily unavailable
code: aiLearning.upstreamFailure
components:
securitySchemes:
bearerAuth:
type: http
scheme: bearer
description: >-
Authenticate every request with a Bearer token. Pass your Reply API key
in the
`Authorization` header:
```
Authorization: Bearer
```
Get your API key from the Reply dashboard: **Settings → API Key**.
````
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.