Lelemon
  • Get started
  • Use cases
  • Guides
  • API basics
  • Webhooks
  • SDK, CLI and MCP
  • API reference
  • Changelog
  • v1 (legacy)
Information
Me
    Retrieve the current keyget
API keys
    List scopesget
Agents
    List agentsgetCreate an agentpostRetrieve an agentgetDelete an agentdeleteUpdate an agentpatchActivate an agentpostPause an agentpostList versionsgetRetrieve a versiongetRestore a versionpostList the fixed facts of an agentgetSave a fixed fact of an agentputForget a fixed fact of an agentdelete
Models and templates
    List agent templatesgetList modelsget
Subaccounts
    List subaccountsgetCreate a subaccountpostRetrieve a subaccountgetDelete a subaccountdeleteUpdate a subaccountpatchGive a subaccount its own keypost
WhatsApp
    List WhatsApp connectionsgetCreate a WhatsApp connection linkpostRetrieve a WhatsApp connectiongetRevoke a pending WhatsApp connectiondeleteList WhatsApp phone numbersgetRetrieve a WhatsApp phone numbergetDisconnect a WhatsApp phone numberdeleteChange the agent, label or state of a numberpatchRead the cart and catalog settings of a numbergetTurn the cart button or the catalog icon of a number on or offpatch
Actions
    List actionsgetCreate an actionpostRead the current authorization textgetRotate the signing secret of your actionspostRetrieve the signing secret of an actiongetRetrieve an actiongetDelete an actiondeleteUpdate an actionpatchTest an actionpostRead the approval floor of a shared actiongetSet the approval floor of a shared actionput
MCP servers
    List MCP serversgetConnect an MCP serverpostRetrieve an MCP servergetDisconnect an MCP serverdeleteUpdate an MCP serverpatchList the tools of an MCP servergetDiscover toolspostRead the run identity of an MCP servergetTurn on or rotate the run identity of an MCP serverpostTurn off the run identity of an MCP serverdelete
Data connectors
    Get the lelemon-connector/v1 formatgetValidate a connector documentpostList data connectorsgetRegister a data connectorpostRetrieve a data connectorgetDelete a data connectordeleteUpdate a data connectorpatchRetrieve the signing secret of a data connectorgetRotate the signing secret of a data connectorpostCheck a data connectorpostRetrieve the manifest of a data connectorgetSearch accounts by emailgetRead the card of an account livegetReveal one masked value of an accountpostGet the link of an account actionpostRead a view livegetList who looked at what through a connectorget
Skills
    List skillsgetCreate a skillpostRetrieve a skillgetDelete a skilldeleteUpdate a skillpatch
Agent tools
    List an agent's toolsgetGrant a toolputRemove a tooldelete
Approvals
    List approvalsgetRetrieve an approvalgetApprove an actionpostReject an actionpostAsk for a human review of a decisionpost
Automated decisions
    List automated decisionsgetRetrieve an automated decisiongetAsk for a human review of an automated decisionpost
Team questions
    List team questionsgetRetrieve a team questiongetRead a team question with its contextgetAnswer a team questionpost
Runs
    Give an agent a jobpostList runsgetRetrieve a rungetCancel a runpostAnswer a run that asked somethingpost
Schedules
    Schedule a job for an agentpostList scheduled tasksgetRetrieve a scheduled taskgetDelete a scheduled taskdeleteChange a scheduled taskpatchFire a scheduled task nowpost
Pipelines
    List pipelinesgetCreate a pipelinepostRetrieve a pipelinegetDelete a pipelinedeleteUpdate a pipelinepatchClone a pipelinepostReorder the stages of a pipelinepost
Stages
    List the stages of a pipelinegetCreate a stagepostRetrieve a stagegetDelete a stagedeleteUpdate a stagepatchMake a stage the entry stagepost
Fields
    List the fields of a pipelinegetCreate a fieldpostRetrieve a fieldgetDelete a fielddeleteUpdate a fieldpatch
Routing rules
    List the routing rules of a pipelinegetCreate a routing rulepostRetrieve a routing rulegetDelete a routing ruledeleteUpdate a routing rulepatch
Knowledge
    List collectionsgetCreate a collectionpostSearch the knowledgepostRetrieve a collectiongetDelete a collectiondeleteUpdate a collectionpatchList documentsgetAdd a documentpostRetrieve a documentgetDelete a documentdeleteUpdate a documentpatchLet an agent search a collectionputStop an agent from searching a collectiondelete
Contacts
    List contactsgetCreate a contactpostCreate or update a contact by your own idputRetrieve a contactgetErase a contactdeleteUpdate a contactpatch
Contact fields
    List the contact fieldsgetCreate a contact fieldpostRetrieve a contact fieldgetDelete a contact fielddeleteUpdate a contact fieldpatch
Companies
    List companiesgetCreate a companypostRetrieve a companygetDelete a companydeleteUpdate a companypatch
Lead forms
    List lead formsgetCreate a lead formpostRetrieve a lead formgetDelete a lead formdeleteUpdate a lead formpatch
Suppressions
    Retrieve a suppressionget
Cases
    List casesgetOpen a casepostRetrieve a casegetDelete a casedeleteUpdate a casepatchMove a case to a stagepostAssign a case to a memberpostClose a casepostReopen a casepostTake control of a casepostHand a case back to the agentpostMark a case as readpostMove several cases to a stagepostArchive a casepostBring an archived case backpostMute the contact of a casepostUnmute the contact of a casepostHand a case to a personpostReveal a sensitive field of a casepostEscalate a case to the technical teampostList the escalations of a caseget
Case queue
    Read the attention queue of a memberget
Contact events
    List the events of a contactgetReport a contact eventpostRetrieve a contact eventget
Case reminders
    List the reminders of a casegetAdd a reminder to a casepost
Reminders
    Retrieve a remindergetDelete a reminderdeleteEdit a reminderpatchComplete a reminderpostDismiss a reminderpost
Presence
    Report who is online in your apppostRead the presence settinggetTurn presence on or offput
Organization
    Get the organizationgetRename the organizationpatchGet the organization settingsgetChange the organization settingspatchGet the company profilegetChange the company profilepatch
Case notes
    List the notes of a casegetAdd a note to a casepostRetrieve a notegetDelete a notedeleteEdit a notepatch
Case attachments
    List the files of a casegetUpload a file to a casepostRetrieve a file of a casegetDelete a file of a casedeleteGet the download link of a fileget
Files
    List filesgetUpload a filepostRetrieve a filegetDelete a filedelete
Case messages
    List the messages of a casegetAdd a message to a casepostRedact a messagepost
Case turns
    List the turns the agent took in a caseget
Usage
    Get the usage of your agentsget
Case metrics
    List the times of the cases that changed in a rangeget
Budgets
    List the budgetsgetCreate a monthly budgetpostGet a budgetgetDelete a budgetdeleteChange a budgetpatch
Members
    List the active members of the organizationgetList who invited, changed the role of or removed membersgetRead whether a team member is availablegetMark a team member available or awayputChange the role of a memberpatchChange which cases a member seespatchRemove a member from the organizationdelete
Invitations
    List the invitations of the organizationgetInvite a person to the organizationpostResend an invitationpostCancel an invitationpost
Scheduling
    List the agendasgetCreate an agendapostRetrieve an agendagetUpdate an agendapatchList the team of an agendagetAdd a person to the team of an agendapostRemove a person from the team of an agendadeleteUpdate a person in the team of an agendapatchRead the free slots of an agendagetList meetingsgetBook a meetingpostRetrieve a meetinggetCancel a meetingpostMove a meeting to another slotpost
Webhook endpoints
    List webhook endpointsgetCreate a webhook endpointpostRetrieve a webhook endpointgetDelete a webhook endpointdeleteUpdate a webhook endpointpatchRotate the signing secretpostSend a test eventpostList the deliveries of an endpointget
Events
    Resend an eventpostList the event typesget
Recipient allowlist
    Get the recipient allowlistgetReplace the recipient allowlistput
WhatsApp templates
    List templatesgetCreate a templatepostSync template statusespostRetrieve a templategetDelete a templatedeleteUpdate a templatepatchRender a templatepostSend a test messagepost
WhatsApp messages
    Send a templatepostRetrieve a template messagegetCancel a scheduled messagepost
Triggers
    List triggersgetCreate a triggerpostRetrieve a triggergetDelete a triggerdeleteUpdate a triggerpatchList the runs of a triggerget
Opt-outs
    List opt-outsgetOpt a contact outputClear an opt-outdelete
Campaigns
    List campaignsgetCreate a campaignpostRetrieve a campaigngetDelete a campaigndeleteUpdate a campaignpatchPreview a campaignpostLaunch a campaignpostPause a campaignpostResume a campaignpostCancel a campaignpostList recipientsget
Email domains
    List the sender domainsgetAdd a sender domainpostGet a sender domain with its status and DNS recordsgetRemove a sender domaindeleteChange the default sender or the inbound agent of a domainpatchCheck the DNS records nowpost
Email messages
    Write first by emailpostRetrieve an email messageget
Teams
    List the teamsgetCreate a teampostRetrieve a teamgetDelete a teamdeleteUpdate a teampatchMerge another team into this onepostMove a pipeline into a teamputAdd a member to a team or change their role in itputRemove a member from a teamdelete
Linear
    Get the Linear connectiongetConnect LinearputDisconnect LineardeleteUpdate the Linear connectionpatch
Schemas
powered by Zudoku
Lelemon API
Lelemon API

Approvals


List approvals

GET
https://api.lelemon.cl
/v2/approvals
Bearer (API key)

Requests an agent made to run an action that its tool policy puts behind a person: refunds, credits, writes. Newest first. Filter by status=pending to build a queue of what waits for a decision. Read only and safe to repeat. To change what needs approval, grant the tool with an approval policy: PUT /v2/agents/{id}/tools/{toolId}.

Required scope: approvals:read.

List approvals › query Parameters

limit
​integer · min: 1 · max: 100
Default: 20
cursor
​string · maxLength: 512

The nextCursor of the previous page, with the same filters.

status
​string · enum
Enum values:
pending
approved
rejected
expired
canceled
executing
executed
failed
agentId
​string
caseId
​string

List approvals › Headers

Lelemon-Subaccount
​string · pattern: ^sub_[0-9A-Za-z]{22}…

The subaccount (sub_…) this request acts on. Only for the keys of a platform account; without it the request acts on the account itself.

List approvals › Responses

ApprovalList
object
​string · enum · required
Enum values:
list
hasMore
​boolean · required
nextCursor
​string | null · required
​Approval[] · required
GET/v2/approvals
curl https://api.lelemon.cl/v2/approvals \ --header 'Authorization: Bearer <token>'
Example Responses
{ "object": "list", "hasMore": true, "nextCursor": "nextCursor", "data": [ { "object": "approval", "id": "apr_5RvP2aKq9XwLmN3tYbZc01", "status": "pending", "agentId": "agt_5RvP2aKq9XwLmN3tYbZc01", "caseId": "cse_5RvP2aKq9XwLmN3tYbZc01", "tool": "issue_refund", "arguments": {}, "summary": "summary", "reasonCode": "policy_always", "ruleIndex": 0, "automatedDecision": true, "dataSeen": [ "string" ], "approvers": [ { "type": "role", "value": "admin" } ], "requestedByKeyId": "requestedByKeyId", "decision": { "decidedAt": "2024-08-25T15:00:00Z", "by": {}, "reason": "reason" }, "result": "result", "executedAt": "2024-08-25T15:00:00Z", "expiresAt": "2024-08-25T15:00:00Z", "escalatedAt": "2024-08-25T15:00:00Z", "reviewRequestedAt": "2024-08-25T15:00:00Z", "reviewReason": "reviewReason", "createdAt": "2024-08-25T15:00:00Z", "updatedAt": "2024-08-25T15:00:00Z" } ] }
json
application/json

Retrieve an approval

GET
https://api.lelemon.cl
/v2/approvals/{id}
Bearer (API key)

The exact arguments the action will run with, why it needed approval, who may decide it, who decided and why, and what the action returned. Use it before deciding, and to answer a claim afterwards. Read only and safe to repeat.

Required scope: approvals:read.

Retrieve an approval › path Parameters

id
​string · pattern: ^apr_[0-9A-Za-z]{22}… · required

The id of the approval (apr_…).

Retrieve an approval › Headers

Lelemon-Subaccount
​string · pattern: ^sub_[0-9A-Za-z]{22}…

The subaccount (sub_…) this request acts on. Only for the keys of a platform account; without it the request acts on the account itself.

Retrieve an approval › Responses

Approval
object
​string · enum · required
Enum values:
approval
id
​string · required
status
​string · enum · required

pending waits for a decision. approved, executing, executed and failed follow an approval: the action runs only after it. rejected, expired and canceled never run it. preview only appears in the answer of approve with dryRun=true: the approval as it is now, with the decision that would be recorded; nothing was decided or run.

Enum values:
pending
approved
rejected
expired
canceled
executing
executed
failed
agentId
​string · required
caseId
​string | null · required
tool
​string · required

The name of the action or tool the agent wants to run.

​object | null · required

The exact arguments the action will run with if approved. Null when the tool is noPersist: those are kept encrypted and shown to nobody.

summary
​string · required

A line for a person to read. For a noPersist tool it is only the name.

reasonCode
​string | null · enum · required

Why it needed approval: policy_always, policy_condition (the condition on its arguments held), policy_accumulated (the sum for this contact in the period passed the limit) or policy_rule (a rule of the policy asked for it: see ruleIndex).

Enum values:
policy_always
policy_condition
policy_accumulated
policy_rule
policy_not_required
ruleIndex
​integer | null · required

The position (from 0) of the rule of the policy that decided this, when reasonCode is policy_rule. Null otherwise.

automatedDecision
​boolean · required

True when the policy marks this as an automated decision about a person (Ley 21.719, art. 8 bis). The person may ask for a review.

dataSeen
​string[] · required

The names of the argument fields the approver was shown. Never their values.

​Approver[] · required
requestedByKeyId
​string | null · required

The API key whose turn proposed the action, if any. It can never decide this request.

​required
result
​string | null · required

What the action returned once it ran. Truncated.

executedAt
​string | null · date-time · required
expiresAt
​string · date-time · required
escalatedAt
​string | null · date-time · required
reviewRequestedAt
​string | null · date-time · required
reviewReason
​string | null · required

What the person said when they asked for the review: their point of view on the decision. Null if they gave none or no review was asked.

createdAt
​string · date-time · required
updatedAt
​string · date-time · required
GET/v2/approvals/{id}
curl https://api.lelemon.cl/v2/approvals/:id \ --header 'Authorization: Bearer <token>'
Example Responses
{ "object": "approval", "id": "apr_5RvP2aKq9XwLmN3tYbZc01", "status": "pending", "agentId": "agt_5RvP2aKq9XwLmN3tYbZc01", "caseId": "cse_5RvP2aKq9XwLmN3tYbZc01", "tool": "issue_refund", "arguments": {}, "summary": "summary", "reasonCode": "policy_always", "ruleIndex": 0, "automatedDecision": true, "dataSeen": [ "string" ], "approvers": [ { "type": "role", "value": "admin" } ], "requestedByKeyId": "requestedByKeyId", "decision": { "decidedAt": "2024-08-25T15:00:00Z", "by": {}, "reason": "reason" }, "result": "result", "executedAt": "2024-08-25T15:00:00Z", "expiresAt": "2024-08-25T15:00:00Z", "escalatedAt": "2024-08-25T15:00:00Z", "reviewRequestedAt": "2024-08-25T15:00:00Z", "reviewReason": "reviewReason", "createdAt": "2024-08-25T15:00:00Z", "updatedAt": "2024-08-25T15:00:00Z" }
json
application/json

Approve an action

POST
https://api.lelemon.cl
/v2/approvals/{id}/approve
Bearer (API key)

Approves the request and runs the action now, with the arguments you read in the approval; then the agent resumes the conversation with the result. This has the real effect of the action: it can move money or write to a person. It works only if the tool policy names this key as approver, the key has the approvals:write scope and the request was not created by this same key. Not idempotent: a request is decided once, and a second decision answers 409 approval_not_pending. A reason is required. Use dryRun=true to run the same checks (reason, key allowed to decide, still pending) and get the approval as it would end, without deciding or running anything.

Required scope: approvals:write.

Approve an action › path Parameters

id
​string · pattern: ^apr_[0-9A-Za-z]{22}… · required

The id of the approval (apr_…).

Approve an action › query Parameters

dryRun
​boolean

When true, the request goes through the same validation as the real one and is answered with what would happen (always 200), and nothing is saved, sent or charged. The answer has the same shape as the real one plus dryRun: true; an id in it is not saved. It does not use the Idempotency-Key.

Approve an action › Headers

Lelemon-Subaccount
​string · pattern: ^sub_[0-9A-Za-z]{22}…

The subaccount (sub_…) this request acts on. Only for the keys of a platform account; without it the request acts on the account itself.

Idempotency-Key
​string · minLength: 1 · maxLength: 255

Makes the request safe to retry. Once it starts running, its final response (2xx, 4xx or 5xx) is kept for 24 hours and the same key gets it back, without repeating the effect. A request rejected before it runs leaves the key free. A replayed create omits the secret. The same key with a different request is rejected with 409 idempotency_key_reused.

Approve an action › Request Body

DecideApproval
reason
​string · minLength: 1 · maxLength: 1000 · pattern: ^[^\u0000]*$ · required

Why you decide this. Required for approving and for rejecting. It is stored with the decision and read by the auditor.

​object

The person in your system who decided. Lelemon records it as declared by this key, not verified. A key cannot declare a role above its own.

Approve an action › Responses

What a dryRun answer adds to the normal one: nothing was saved, sent or charged, and an id in it is not saved.
object
​string · enum · required
Enum values:
approval
id
​string · required
status
​string · enum · required

pending waits for a decision. approved, executing, executed and failed follow an approval: the action runs only after it. rejected, expired and canceled never run it. preview only appears in the answer of approve with dryRun=true: the approval as it is now, with the decision that would be recorded; nothing was decided or run.

Enum values:
pending
approved
rejected
expired
canceled
executing
executed
failed
agentId
​string · required
caseId
​string | null · required
tool
​string · required

The name of the action or tool the agent wants to run.

​object | null · required

The exact arguments the action will run with if approved. Null when the tool is noPersist: those are kept encrypted and shown to nobody.

summary
​string · required

A line for a person to read. For a noPersist tool it is only the name.

reasonCode
​string | null · enum · required

Why it needed approval: policy_always, policy_condition (the condition on its arguments held), policy_accumulated (the sum for this contact in the period passed the limit) or policy_rule (a rule of the policy asked for it: see ruleIndex).

Enum values:
policy_always
policy_condition
policy_accumulated
policy_rule
policy_not_required
ruleIndex
​integer | null · required

The position (from 0) of the rule of the policy that decided this, when reasonCode is policy_rule. Null otherwise.

automatedDecision
​boolean · required

True when the policy marks this as an automated decision about a person (Ley 21.719, art. 8 bis). The person may ask for a review.

dataSeen
​string[] · required

The names of the argument fields the approver was shown. Never their values.

​Approver[] · required
requestedByKeyId
​string | null · required

The API key whose turn proposed the action, if any. It can never decide this request.

​required
result
​string | null · required

What the action returned once it ran. Truncated.

executedAt
​string | null · date-time · required
expiresAt
​string · date-time · required
escalatedAt
​string | null · date-time · required
reviewRequestedAt
​string | null · date-time · required
reviewReason
​string | null · required

What the person said when they asked for the review: their point of view on the decision. Null if they gave none or no review was asked.

createdAt
​string · date-time · required
updatedAt
​string · date-time · required
dryRun
​boolean · const

Present and true when the request had dryRun=true.

Const value:
POST/v2/approvals/{id}/approve
curl https://api.lelemon.cl/v2/approvals/:id/approve \ --request POST \ --header 'Content-Type: application/json' \ --header 'Authorization: Bearer <token>' \ --data '{ "reason": "Within the refund policy: order delivered damaged.", "actor": { "id": "user_8841", "role": "staff" } }'
Example Request Body
{ "reason": "Within the refund policy: order delivered damaged.", "actor": { "id": "user_8841", "role": "staff" } }
json
application/json
Example Responses
{ "object": "approval", "id": "apr_5RvP2aKq9XwLmN3tYbZc01", "status": "pending", "agentId": "agt_5RvP2aKq9XwLmN3tYbZc01", "caseId": "cse_5RvP2aKq9XwLmN3tYbZc01", "tool": "issue_refund", "arguments": {}, "summary": "summary", "reasonCode": "policy_always", "ruleIndex": 0, "automatedDecision": true, "dataSeen": [ "string" ], "approvers": [ { "type": "role", "value": "admin" } ], "requestedByKeyId": "requestedByKeyId", "decision": { "decidedAt": "2024-08-25T15:00:00Z", "by": {}, "reason": "reason" }, "result": "result", "executedAt": "2024-08-25T15:00:00Z", "expiresAt": "2024-08-25T15:00:00Z", "escalatedAt": "2024-08-25T15:00:00Z", "reviewRequestedAt": "2024-08-25T15:00:00Z", "reviewReason": "reviewReason", "createdAt": "2024-08-25T15:00:00Z", "updatedAt": "2024-08-25T15:00:00Z", "dryRun": true }
json
application/json

Reject an action

POST
https://api.lelemon.cl
/v2/approvals/{id}/reject
Bearer (API key)

Rejects the request: the action never runs and the agent tells the person it was not done. Same rules as approve: the policy must name this key, and a request is decided once (409 approval_not_pending after that). A reason is required.

Required scope: approvals:write.

Reject an action › path Parameters

id
​string · pattern: ^apr_[0-9A-Za-z]{22}… · required

The id of the approval (apr_…).

Reject an action › Headers

Lelemon-Subaccount
​string · pattern: ^sub_[0-9A-Za-z]{22}…

The subaccount (sub_…) this request acts on. Only for the keys of a platform account; without it the request acts on the account itself.

Idempotency-Key
​string · minLength: 1 · maxLength: 255

Makes the request safe to retry. Once it starts running, its final response (2xx, 4xx or 5xx) is kept for 24 hours and the same key gets it back, without repeating the effect. A request rejected before it runs leaves the key free. A replayed create omits the secret. The same key with a different request is rejected with 409 idempotency_key_reused.

Reject an action › Request Body

DecideApproval
reason
​string · minLength: 1 · maxLength: 1000 · pattern: ^[^\u0000]*$ · required

Why you decide this. Required for approving and for rejecting. It is stored with the decision and read by the auditor.

​object

The person in your system who decided. Lelemon records it as declared by this key, not verified. A key cannot declare a role above its own.

Reject an action › Responses

Approval
object
​string · enum · required
Enum values:
approval
id
​string · required
status
​string · enum · required

pending waits for a decision. approved, executing, executed and failed follow an approval: the action runs only after it. rejected, expired and canceled never run it. preview only appears in the answer of approve with dryRun=true: the approval as it is now, with the decision that would be recorded; nothing was decided or run.

Enum values:
pending
approved
rejected
expired
canceled
executing
executed
failed
agentId
​string · required
caseId
​string | null · required
tool
​string · required

The name of the action or tool the agent wants to run.

​object | null · required

The exact arguments the action will run with if approved. Null when the tool is noPersist: those are kept encrypted and shown to nobody.

summary
​string · required

A line for a person to read. For a noPersist tool it is only the name.

reasonCode
​string | null · enum · required

Why it needed approval: policy_always, policy_condition (the condition on its arguments held), policy_accumulated (the sum for this contact in the period passed the limit) or policy_rule (a rule of the policy asked for it: see ruleIndex).

Enum values:
policy_always
policy_condition
policy_accumulated
policy_rule
policy_not_required
ruleIndex
​integer | null · required

The position (from 0) of the rule of the policy that decided this, when reasonCode is policy_rule. Null otherwise.

automatedDecision
​boolean · required

True when the policy marks this as an automated decision about a person (Ley 21.719, art. 8 bis). The person may ask for a review.

dataSeen
​string[] · required

The names of the argument fields the approver was shown. Never their values.

​Approver[] · required
requestedByKeyId
​string | null · required

The API key whose turn proposed the action, if any. It can never decide this request.

​required
result
​string | null · required

What the action returned once it ran. Truncated.

executedAt
​string | null · date-time · required
expiresAt
​string · date-time · required
escalatedAt
​string | null · date-time · required
reviewRequestedAt
​string | null · date-time · required
reviewReason
​string | null · required

What the person said when they asked for the review: their point of view on the decision. Null if they gave none or no review was asked.

createdAt
​string · date-time · required
updatedAt
​string · date-time · required
POST/v2/approvals/{id}/reject
curl https://api.lelemon.cl/v2/approvals/:id/reject \ --request POST \ --header 'Content-Type: application/json' \ --header 'Authorization: Bearer <token>' \ --data '{ "reason": "Above the limit of the refund policy.", "actor": { "id": "user_8841", "role": "staff" } }'
Example Request Body
{ "reason": "Above the limit of the refund policy.", "actor": { "id": "user_8841", "role": "staff" } }
json
application/json
Example Responses
{ "object": "approval", "id": "apr_5RvP2aKq9XwLmN3tYbZc01", "status": "pending", "agentId": "agt_5RvP2aKq9XwLmN3tYbZc01", "caseId": "cse_5RvP2aKq9XwLmN3tYbZc01", "tool": "issue_refund", "arguments": {}, "summary": "summary", "reasonCode": "policy_always", "ruleIndex": 0, "automatedDecision": true, "dataSeen": [ "string" ], "approvers": [ { "type": "role", "value": "admin" } ], "requestedByKeyId": "requestedByKeyId", "decision": { "decidedAt": "2024-08-25T15:00:00Z", "by": {}, "reason": "reason" }, "result": "result", "executedAt": "2024-08-25T15:00:00Z", "expiresAt": "2024-08-25T15:00:00Z", "escalatedAt": "2024-08-25T15:00:00Z", "reviewRequestedAt": "2024-08-25T15:00:00Z", "reviewReason": "reviewReason", "createdAt": "2024-08-25T15:00:00Z", "updatedAt": "2024-08-25T15:00:00Z" }
json
application/json

Ask for a human review of a decision

POST
https://api.lelemon.cl
/v2/approvals/{id}/request-review
Bearer (API key)

Use it when the person affected by a decision asks for it to be reviewed (Ley 21.719, art. 8 bis). It works only on a decided approval whose policy has automatedDecision. It fires the approval.review_requested webhook so your team can act, and it changes nothing else. Idempotent: asking twice fires the event once.

Required scope: approvals:write.

Ask for a human review of a decision › path Parameters

id
​string · pattern: ^apr_[0-9A-Za-z]{22}… · required

The id of the approval (apr_…).

Ask for a human review of a decision › Headers

Lelemon-Subaccount
​string · pattern: ^sub_[0-9A-Za-z]{22}…

The subaccount (sub_…) this request acts on. Only for the keys of a platform account; without it the request acts on the account itself.

Idempotency-Key
​string · minLength: 1 · maxLength: 255

Makes the request safe to retry. Once it starts running, its final response (2xx, 4xx or 5xx) is kept for 24 hours and the same key gets it back, without repeating the effect. A request rejected before it runs leaves the key free. A replayed create omits the secret. The same key with a different request is rejected with 409 idempotency_key_reused.

Ask for a human review of a decision › Request Body optional

RequestApprovalReview
reason
​string · maxLength: 1000

The point of view of the person on the decision, in their words. It is stored with the record, shown in reviewReason and sent in the review_requested event. Do not put anything the person did not say.

Ask for a human review of a decision › Responses

Approval
object
​string · enum · required
Enum values:
approval
id
​string · required
status
​string · enum · required

pending waits for a decision. approved, executing, executed and failed follow an approval: the action runs only after it. rejected, expired and canceled never run it. preview only appears in the answer of approve with dryRun=true: the approval as it is now, with the decision that would be recorded; nothing was decided or run.

Enum values:
pending
approved
rejected
expired
canceled
executing
executed
failed
agentId
​string · required
caseId
​string | null · required
tool
​string · required

The name of the action or tool the agent wants to run.

​object | null · required

The exact arguments the action will run with if approved. Null when the tool is noPersist: those are kept encrypted and shown to nobody.

summary
​string · required

A line for a person to read. For a noPersist tool it is only the name.

reasonCode
​string | null · enum · required

Why it needed approval: policy_always, policy_condition (the condition on its arguments held), policy_accumulated (the sum for this contact in the period passed the limit) or policy_rule (a rule of the policy asked for it: see ruleIndex).

Enum values:
policy_always
policy_condition
policy_accumulated
policy_rule
policy_not_required
ruleIndex
​integer | null · required

The position (from 0) of the rule of the policy that decided this, when reasonCode is policy_rule. Null otherwise.

automatedDecision
​boolean · required

True when the policy marks this as an automated decision about a person (Ley 21.719, art. 8 bis). The person may ask for a review.

dataSeen
​string[] · required

The names of the argument fields the approver was shown. Never their values.

​Approver[] · required
requestedByKeyId
​string | null · required

The API key whose turn proposed the action, if any. It can never decide this request.

​required
result
​string | null · required

What the action returned once it ran. Truncated.

executedAt
​string | null · date-time · required
expiresAt
​string · date-time · required
escalatedAt
​string | null · date-time · required
reviewRequestedAt
​string | null · date-time · required
reviewReason
​string | null · required

What the person said when they asked for the review: their point of view on the decision. Null if they gave none or no review was asked.

createdAt
​string · date-time · required
updatedAt
​string · date-time · required
POST/v2/approvals/{id}/request-review
curl https://api.lelemon.cl/v2/approvals/:id/request-review \ --request POST \ --header 'Content-Type: application/json' \ --header 'Authorization: Bearer <token>' \ --data '{ "reason": "The customer disagrees with the decision." }'
Example Request Body
{ "reason": "The customer disagrees with the decision." }
json
application/json
Example Responses
{ "object": "approval", "id": "apr_5RvP2aKq9XwLmN3tYbZc01", "status": "pending", "agentId": "agt_5RvP2aKq9XwLmN3tYbZc01", "caseId": "cse_5RvP2aKq9XwLmN3tYbZc01", "tool": "issue_refund", "arguments": {}, "summary": "summary", "reasonCode": "policy_always", "ruleIndex": 0, "automatedDecision": true, "dataSeen": [ "string" ], "approvers": [ { "type": "role", "value": "admin" } ], "requestedByKeyId": "requestedByKeyId", "decision": { "decidedAt": "2024-08-25T15:00:00Z", "by": {}, "reason": "reason" }, "result": "result", "executedAt": "2024-08-25T15:00:00Z", "expiresAt": "2024-08-25T15:00:00Z", "escalatedAt": "2024-08-25T15:00:00Z", "reviewRequestedAt": "2024-08-25T15:00:00Z", "reviewReason": "reviewReason", "createdAt": "2024-08-25T15:00:00Z", "updatedAt": "2024-08-25T15:00:00Z" }
json
application/json

Agent toolsAutomated decisions