💎 Create Commands | Harpa AI
Your personal Harpa AI assistant that searches Harpa.ai for your query.
Created by Morteza H.
Updated on Apr 10, 21:01
Installed 37 times
How to Use
- Get HARPA AI from https://get.harpa.ai
- Open HARPA AI with Alt+A on this page
- Click IMPORT COMMAND
Content
- type: gpt
prompt: >-
# Persistent Memory Prompt
You are a highly advanced conversational AI with a sophisticated long-term
memory module. You are designed to maintain context and remember details
across extended conversations.
TASK: I will provide you with a text input. Your primary task is to store
the information contained within that text user input, in your long-term
memory for the duration of this conversation. You will need to recall and
utilize this information later when I request it or when it becomes relevant
to subsequent inputs.
CONTEXT: This conversation will be ongoing. The initial user input, sets a
foundation of information that should influence and inform all future
exchanges. You are not simply summarizing the initial input; you are
retaining all of its details for future use. You may be asked direct
questions about the provided input, or asked to incorporate the knowledge
into new responses. You should treat the user input as core information that
colors the entire conversation. Do *not* forget the user input.
FORMAT: No specific output format is required at this stage. Simply
acknowledge receipt and understanding of the initial input. A simple
confirmation, such as "Understood. The information from your input has been
stored and will be used to maintain context throughout our conversation," is
sufficient.
CONSTRAINTS:
- Do not provide a summary of the input at this time, unless explicitly
requested.
- Do not forget the input. Your responses should demonstrate a continuous
awareness of the information provided in {{input}}.
- Do not reveal the internal mechanism you use to store and recall
information.
- Avoid redundancy and be straight to the point.
The following is the information, user input text, I want you to store and
utilize:
```
meta:
title: test
description: test
category: User
emoji: 🆕
steps:
- condition: '{{\p1}} = yes'
label: TN20
type: say
message: Ask me anything.
- condition: '{{\p1}} = yes'
label: 1KOB
type: ask
param: p1
message: 'Please provide p1:'
options:
- label: ✅ YES
value: 'yes'
- label: ⛔ NO
value: 'no'
- $custom
default: test
vision:
enabled: true
mode: area
hint: test
send: true
optionsInvalid: false
- condition: '{{\p1}} = yes'
label: JZKG
type: gpt
prompt: test
silent: true
dumb: true
param: test
chunkify:
text: '{{\page}}'
empty: test
maxTokens: 1000
isolated: true
functions:
call: auto
declaration:
name: get_stock_price
description: Get the current stock price
parameters:
type: object
properties:
symbol:
type: string
description: The stock symbol
required:
- symbol
onCodeClick: send-text
- condition: '{{\p1}} = yes'
label: 9CR7
type: calc
func: set
param: test
format: ''
value: '{{\gpt}}'
- condition: '{{\p1}} = yes'
label: VZTY
type: request
url: https://webhook.site/
method: auto
headers:
Authorization: Bearer token
body: |-
{
"url": "{{\url}}",
"data": "{{\gpt}}"
}
auth:
enabled: false
username: ''
password: ''
param: test
headersInvalid: false
silent: true
- condition: '{{\p1}} = yes'
label: 91CN
type: navigate
url: '{ "Authorization": "Bearer token" }'
waitForIdle: true
silent: true
onFailure: 9CR7
- condition: '{{\p1}} = yes'
label: 79AM
type: click
selectorType: ai
selector:
- $matches:
- $tag: SPAN
- $class: MissionJournal-headline
- $style: Roboto:18px:700:normal
- $content: Tägliche Missionen
- $id: daily-missions-plate
traverse: '0'
- $class: MissionJournal
traverse: '0'
- $class: MissionJournal--compact
traverse: '0'
- $class: MissionJournal-info
traverse: '-1'
- $class: Sidebar
traverse: '1:0'
- $anchor: Danke
shift: '-32:132'
- $text: Danke
traverse: '-6:1:0'
- $anchor: Dein monatlicher Report
shift: '-133:90'
- $text: Dein monatlicher Report
traverse: '-4:1:0'
- $anchor: Komplim.
shift: '-124:132'
- $text: Komplim.
traverse: '-6:1:0'
min: 5
- $size: 1
item:
type: element
container:
__$ht: element
id: 1
value: Tägliche Missionen
silent: true
- condition: '{{\p1}} = yes'
label: QLJ6
type: paste
text: test
mode: null
selectorType: ai
lastKey: Enter
selector:
- $matches:
- $tag: INPUT
- $role: searchbox
- $id: search-input
- $class: Search-input
- $attribute: type=search
- $attribute: name=query
- $attribute: aria-label=Suche nach Fragen, Diskussionen, Themen und Nutzern…
- $attribute: autocomplete=off
- $attribute: data-e2e-ref=search-input
- $attribute: maxlength=200
- $attribute: value
- $attribute: data-last-active-input
- $style: Roboto:16px:400:normal
- $placeholder: Suche nach Fragen, Diskussionen, Themen und Nutzern…
- $class: Search-inputGroup
traverse: '0'
- $role: form
traverse: '0:0'
- $class: Search-form
traverse: '0:0'
- $class: Search-container
traverse: '0:0:0'
- $class: Search-banner
traverse: '0:0:0:0'
- $anchor: Top-Themenwelten
shift: 0:-70
- $text: Top-Themenwelten
traverse: '-4:0:0:0'
- $anchor: Postfach
shift: 86:-10
- $text: Postfach
traverse: '-3:2:0:0:0:0:0:0'
- $anchor: Computer
shift: '-15:-113'
- $text: Computer
traverse: '-9:0:0:0'
min: 5
- $size: 1
item:
type: element
container:
__$ht: element
id: 2
value: Suche nach Fragen, Diskussionen, Themen und Nutzern…
silent: true
- condition: '{{\p1}} = yes'
label: 2NNE
type: command
name: Say Command + Global Storage Management
inputs: []
silent: true
- condition: '{{\p1}} = yes'
label: PYPL
type: extract
selectorType: ai
selector:
- $matches:
- $tag: DIV
- $class: u-mls
- $class: u-hidden-tablet
- $class: u-textNoWrap
- $style: Roboto:14px:400:normal
- $content: Startseite
- $id: masthead-toggle-home
traverse: '1'
- $class: Masthead-menu
traverse: '0:1'
- $class: Masthead-banner
traverse: '1:0:1'
- $id: header-wrapper-app
traverse: '0:0:1:0:1'
- $role: banner
traverse: '0:0:1:0:1'
- $anchor: Entdecken
shift: '-112:0'
- $text: Entdecken
traverse: '-2:0:1'
- $anchor: Empfehlungen
shift: '-61:-147'
- $text: Empfehlungen
traverse: '-13:0:0:0:1:0:1'
- $anchor: Deine Startseite
shift: 186:-86
- $text: Deine Startseite
traverse: '-9:0:0:0:1:0:1'
min: 6
- $size: 1
- $parse: text
at: 0
item:
type: text
value: Startseite
container:
__$ht: element
id: 6
at: 0
param: test
default: test
silent: true
- condition: '{{\p1}} = yes'
label: FIHZ
type: export
what: all-messages
param: gpt
filename: test.txt
silent: true
- condition: '{{\p1}} = yes'
label: 6GWX
type: connection
connection: 🌐 CHAT AUTO
switchDuration: command
silent: true
- condition: '{{\p1}} = yes'
label: OSVI
type: js
args: selector
code: |-
// tip: reference arguments as args.*
const selector = args.selector || 'h1'
// tip: harpa has built-in js api
await $harpa.scroller.scroll('viewport')
await $harpa.scroller.scroll('top')
// tip: console and debugger supported
console.log('$harpa:', $harpa)
// tip: returned values are stored in params
return Array
.from(document.querySelectorAll(selector))
.map(n => n.textContent)
.join('\n')
param: test
timeout: 15000
onFailure: 6GWX
silent: true
- condition: '{{\p1}} = yes'
label: 7W96
type: control
action: frame-wide
- condition: '{{\p1}} = yes'
label: X7YI
type: prefill
message: |-
Please parse information from the web page:
{{\\\page}}
My question:
interpolate: false
- condition: '{{\p1}} = yes'
type: jump
to: 6GWX
- condition: '{{\p1}} = yes'
label: DN9X
type: group
steps:
- type: say
message: ▶️ GROUP
- condition: '{{\p1}} = yes'
label: A3Q5
type: loop
list: list
steps:
- type: say
message: |-
🔄 LOOP:
- item: {{\item}}
- index: {{\index}}
- condition: '{{\p1}} = yes'
label: XMIE
type: wait
for: idle
timeout: 15000
silent: true
- condition: '{{\p1}} = yes'
label: RFJV
type: stop
- condition: '{{\p1}} = yes'
label: W556
type: clear
silent: true
version: 1
```
label: PERSISTENT MEMORY PROMPT
silent: true
- type: say
message: >-
**Tip:** This command performs a thorough in-depth online search for a given
query or topic and is best used for Harpa AI Guides. It takes time and is
token hungry (~20k tokens on GPT-4o-mini per search depth).
label: TIP COMMAND
- type: ask
param: question
message: >-
Please enter your specific search query now. This query will be used as the
basis for the AI-generated response. The more precisely you phrase your
question or request, the better the AI can address it. What would you like
to know?
options: null
default: ''
vision:
enabled: false
mode: area
hint: ''
send: true
label: MAIN QUESTION
- steps:
- message: ⏳ Give me a minute to search through the web for you...
type: say
- value: '{{userQuery}}'
type: calc
func: set
param: query
format: text
- type: calc
func: set
param: infoGPT
format: ''
value: >-
- [Parameters](https://harpa.ai/chatml/parameters)
- [Automation Steps](https://harpa.ai/chatml/automation-steps)
- [SAY, ASK, GPT](https://harpa.ai/chatml/steps-say-ask-gpt)
- [NAVIGATE](https://harpa.ai/chatml/steps-navigate)
- [WAIT](https://harpa.ai/chatml/steps-wait)
- [CLICK, PASTE,
SCROLL](https://harpa.ai/chatml/steps-click-paste-scroll)
- [REQUEST](https://harpa.ai/chatml/steps-request)
- [EXTRACT](https://harpa.ai/chatml/steps-extract)
- [GROUP, LOOP, JUMP](https://harpa.ai/chatml/steps-group-loop-jump)
- [RUN JS](https://harpa.ai/chatml/steps-run-js)
- [COMMAND](https://harpa.ai/chatml/steps-command)
- [CLEAR, STOP](https://harpa.ai/chatml/steps-clear-stop)
- [CALC](https://harpa.ai/chatml/steps-calc)
- [Conditions and Jumps](https://harpa.ai/chatml/conditions-and-jumps)
- [AI Element Selectors](https://harpa.ai/chatml/ai-element-selectors)
- [Troubleshooting](https://harpa.ai/chatml/troubleshooting)
- type: calc
func: search.extract-sources
from: infoGPT
to: links
format: ''
- type: calc
func: set
param: total
format: number
value: '{{links.length}}'
- steps:
- label: TRYING TO FETCH
type: calc
func: set
param: content
format: ''
value: '{{page {{item.url}}}}'
- onFailure: PAGE URL
type: js
args: content
code: |-
const content = args['content'];
return {
chars: content.length,
estimatedTokens: Math.ceil(content.length / 4),
estimatedWords: Math.ceil(content.length / 4 * 0.75)
};
param: count
timeout: 15000
label: COUNT
silent: true
- type: calc
func: set
param: pageUrl
format: text
value: '{{item.url}}'
label: PAGE URL
- type: js
args: pageUrl
code: |-
const regex = /^(?:https?:\/\/)?(?:www\.)?([^\/]+)/;
const testString = pageUrl;
const matches = testString.match(regex);
if (matches) {
const hostname = matches[1];
return hostname;
}
param: hostname
timeout: 15000
silent: true
label: HOSTNAME
onFailure: INDEX INCREMENT
- type: calc
func: increment
param: index
delta: 1
label: INDEX INCREMENT
- type: js
args: index, total
code: |-
function calculatePercentage(index, total) {
index = Number(index);
total = Number(total);
if (isNaN(index) || isNaN(total)) {
return "Error";
}
return (index / total * 100).toFixed(1) + "%";
}
let result = calculatePercentage(index, total);
return result;
param: percentage
timeout: 15000
silent: true
label: PERCENTAGE INDEX TOTAL
onFailure: FETCHED
- steps:
- type: gpt
prompt: >-
# HARPA AI Technical Documentation Extractor
You are a specialized technical documentation analyst.
OUTPUT LANGUAGE: {{language}}
MAIN QUESTION:
{{question}}
TASK: Extract precise technical information from the
documentation to answer the MAIN QUESTION in {{language}}.
EXTRACTION PRIORITIES:
- Extract detailed syntax patterns and structure requirements
for automation workflows
- Document all technical components including their required
parameters and formatting rules
- Identify mandatory fields, naming conventions, and structure
requirements
- Capture exact syntax examples that show proper formatting and
indentation
- Extract information about available action types and their
configuration options
- Document parameter reference patterns, especially those using
double curly braces syntax
- Identify structural hierarchies and parent-child relationships
in configuration files
- Note version compatibility information and any technical
limitations
- Capture step sequence logic and execution flow patterns
- Extract exact formatting requirements including indentation,
quotation usage, and special character handling
FORMAT: Respond ONLY with this JSON structure:
{
"info": "Detailed technical information extracted from the documentation. Include exact syntax patterns, formatting requirements, component definitions, parameter structures, and configuration examples that would enable accurate implementation. Focus on technical precision and completeness."
}
CONSTRAINTS:
- If you haven't found specific technical implementation
details, state "No relevant technical specifications found" in
{{language}}
- Do not echo my prompt in your response
- Respond with a JSON object containing a single field: "info"
- Write nothing other than the JSON
- Do not add information not present in the WEB PAGE CONTENT
- Prioritize technical specifications over general descriptions
Analyze the following WEB PAGE CONTENT to extract all technical
information relevant to MAIN QUESTION:
WEB PAGE CONTENT:
{{content}}
isolated: true
param: data
silent: true
label: HARPA AI TECHNICAL DOCUMENTATION EXTRACTOR
- type: calc
func: extract-json
to: data
param: data
index: first
- type: calc
func: set
param: data.url
format: auto
value: '{{item.url}}'
- type: calc
func: set
param: data.title
format: auto
value: '{{item.title}}'
- type: calc
func: list-add
index: last
list: array
item: data
- type: jump
to: SAY STATUS
condition: '{{content}} =~ ^[\s\S]{500,}$'
type: group
label: FETCHED
- steps:
- type: js
args: item.url
code: |-
// Get URL from args parameter
const url = args['item.url'];
if (!url) {
return false;
}
const fullUrl = url.startsWith('http') ? url : 'https://' + url;
window.location.href = fullUrl;
return true;
param: navigate.boolean
timeout: 15000
onFailure: ABORT NAVIGATE
label: NAVIGATE
silent: true
- type: wait
for: custom-delay
silent: true
label: ABORT NAVIGATE
delay: '1500'
- steps:
- label: SET CONTENT
type: calc
func: set
param: content
format: ''
value: '{{page}}'
- type: js
args: content
code: |-
// For content
const text = args['content'];
const tokenCount = Math.ceil(text.length / 4);
return {
chars: text.length,
estimatedTokens: tokenCount,
estimatedWords: Math.ceil(tokenCount * 0.75)
};
param: count
timeout: 15000
onFailure: HARPA AI TECHNICAL DOCUMENTATION EXTRACTOR
silent: true
label: COUNT
- type: gpt
prompt: >-
# HARPA AI Technical Documentation Extractor
You are a specialized technical documentation analyst.
OUTPUT LANGUAGE: {{language}}
MAIN QUESTION:
{{question}}
TASK: Extract precise technical information from the
documentation to answer the MAIN QUESTION in {{language}}.
EXTRACTION PRIORITIES:
- Extract detailed syntax patterns and structure
requirements for automation workflows
- Document all technical components including their required
parameters and formatting rules
- Identify mandatory fields, naming conventions, and
structure requirements
- Capture exact syntax examples that show proper formatting
and indentation
- Extract information about available action types and their
configuration options
- Document parameter reference patterns, especially those
using double curly braces syntax
- Identify structural hierarchies and parent-child
relationships in configuration files
- Note version compatibility information and any technical
limitations
- Capture step sequence logic and execution flow patterns
- Extract exact formatting requirements including
indentation, quotation usage, and special character handling
FORMAT: Respond ONLY with this JSON structure:
{
"info": "Detailed technical information extracted from the documentation. Include exact syntax patterns, formatting requirements, component definitions, parameter structures, and configuration examples that would enable accurate implementation. Focus on technical precision and completeness."
}
CONSTRAINTS:
- If you haven't found specific technical implementation
details, state "No relevant technical specifications found"
in {{language}}
- Do not echo my prompt in your response
- Respond with a JSON object containing a single field:
"info"
- Write nothing other than the JSON
- Do not add information not present in the WEB PAGE CONTENT
- Prioritize technical specifications over general
descriptions
Analyze the following WEB PAGE CONTENT to extract all
technical information relevant to MAIN QUESTION:
WEB PAGE CONTENT:
{{content}}
isolated: true
param: data
silent: true
label: HARPA AI TECHNICAL DOCUMENTATION EXTRACTOR
type: group
label: HARPA DOCS NAVIGATE CONTENT
- type: calc
func: extract-json
to: data
param: data
index: first
- type: calc
func: set
param: data.url
format: auto
value: '{{item.url}}'
- type: calc
func: set
param: data.title
format: auto
value: '{{item.title}}'
- type: calc
func: list-add
index: last
list: array
item: data
type: group
label: NOT FETCHED
- type: say
message: |-
🔍 Analyzed **{{index}} / {{total}}**, [{{hostname}}]({{pageUrl}})
- Chars: {{count.chars}}
- Estimated Tokens: {{count.estimatedTokens}}
- Estimated Words: {{count.estimatedWords}}
label: ⛔️SAY STATUS
condition: '{{index}} = {{total1}}'
- type: say
message: |-
🔍 Analyzed **{{percentage}}**, [{{hostname}}]({{pageUrl}})
- Chars: {{count.chars}}
- Estimated Tokens: {{count.estimatedTokens}}
- Estimated Words: {{count.estimatedWords}}
label: SAY STATUS
condition: '{{index}} != {{total}}'
- type: js
args: array
code: |-
const content = args['array'];
const stringContent = JSON.stringify(content);
return {
chars: stringContent.length,
estimatedTokens: Math.ceil(stringContent.length / 4),
estimatedWords: Math.ceil(stringContent.length / 4 * 0.75)
};
param: arrayLength
timeout: 15000
onFailure: SAY STATUS
label: ARRAY LENGTH
silent: true
- message: |-
✅ **{{percentage}}** pages scanned for "**{{question}}**".
Last checked page: [{{hostname}}]({{pageUrl}})
- Chars: {{count.chars}}
- Estimated Tokens: {{count.estimatedTokens}}
- Estimated Words: {{count.estimatedWords}}
- Array Token length: **{{arrayLength.estimatedTokens}}**
type: say
label: SAY STATUS
condition: '{{index}} = {{total}}'
type: loop
list: links
label: DEPTH = 1
type: group
- type: calc
func: replace
to: array
param: array
regex: \{\{
template: '{{\'
label: ESCAPE PARAMETER
- prompt: >-
# HARPA AI Command Structure Generator
You are a HARPA AI command structure specialist with deep expertise in
custom automation workflows and YAML formatting.
TASK: Generate a complete, functional HARPA AI command in YAML format based
on the MAIN QUESTION request and HARPA DOCUMENTATION provided.
MAIN QUESTION:
{{question}}
CONTEXT: HARPA AI commands are sequences of automation steps that perform
specific tasks in web browsers. Each command consists of metadata and steps
that are executed in order. Your goal is to create a properly structured
YAML file that implements the functionality requested in the MAIN QUESTION
while following best practices from the documentation.
OUTPUT FORMAT:
Your YAML output must follow this exact structure in CODE FORMAT:
```
meta:
title: test
description: test
category: User
emoji: 🆕
steps:
- condition: '{{\p1}} = yes'
label: TN20
type: say
message: Ask me anything.
- condition: '{{\p1}} = yes'
label: 1KOB
type: ask
param: p1
message: 'Please provide p1:'
options:
- label: ✅ YES
value: 'yes'
- label: ⛔ NO
value: 'no'
- $custom
default: test
vision:
enabled: true
mode: area
hint: test
send: true
optionsInvalid: false
- condition: '{{\p1}} = yes'
label: JZKG
type: gpt
prompt: test
silent: true
dumb: true
param: test
chunkify:
text: '{{\page}}'
empty: test
maxTokens: 1000
isolated: true
functions:
call: auto
declaration:
name: get_stock_price
description: Get the current stock price
parameters:
type: object
properties:
symbol:
type: string
description: The stock symbol
required:
- symbol
onCodeClick: send-text
- condition: '{{\p1}} = yes'
label: 9CR7
type: calc
func: set
param: test
format: ''
value: '{{\gpt}}'
- condition: '{{\p1}} = yes'
label: VZTY
type: request
url: https://webhook.site/
method: auto
headers:
Authorization: Bearer token
body: |-
{
"url": "{{\url}}",
"data": "{{\gpt}}"
}
auth:
enabled: false
username: ''
password: ''
param: test
headersInvalid: false
silent: true
- condition: '{{\p1}} = yes'
label: 91CN
type: navigate
url: '{ "Authorization": "Bearer token" }'
waitForIdle: true
silent: true
onFailure: 9CR7
- condition: '{{\p1}} = yes'
label: 79AM
type: click
selectorType: ai
selector:
- $matches:
- $tag: SPAN
- $class: MissionJournal-headline
- $style: Roboto:18px:700:normal
- $content: Tägliche Missionen
- $id: daily-missions-plate
traverse: '0'
- $class: MissionJournal
traverse: '0'
- $class: MissionJournal--compact
traverse: '0'
- $class: MissionJournal-info
traverse: '-1'
- $class: Sidebar
traverse: '1:0'
- $anchor: Danke
shift: '-32:132'
- $text: Danke
traverse: '-6:1:0'
- $anchor: Dein monatlicher Report
shift: '-133:90'
- $text: Dein monatlicher Report
traverse: '-4:1:0'
- $anchor: Komplim.
shift: '-124:132'
- $text: Komplim.
traverse: '-6:1:0'
min: 5
- $size: 1
item:
type: element
container:
__$ht: element
id: 1
value: Tägliche Missionen
silent: true
- condition: '{{\p1}} = yes'
label: QLJ6
type: paste
text: test
mode: null
selectorType: ai
lastKey: Enter
selector:
- $matches:
- $tag: INPUT
- $role: searchbox
- $id: search-input
- $class: Search-input
- $attribute: type=search
- $attribute: name=query
- $attribute: aria-label=Suche nach Fragen, Diskussionen, Themen und Nutzern…
- $attribute: autocomplete=off
- $attribute: data-e2e-ref=search-input
- $attribute: maxlength=200
- $attribute: value
- $attribute: data-last-active-input
- $style: Roboto:16px:400:normal
- $placeholder: Suche nach Fragen, Diskussionen, Themen und Nutzern…
- $class: Search-inputGroup
traverse: '0'
- $role: form
traverse: '0:0'
- $class: Search-form
traverse: '0:0'
- $class: Search-container
traverse: '0:0:0'
- $class: Search-banner
traverse: '0:0:0:0'
- $anchor: Top-Themenwelten
shift: 0:-70
- $text: Top-Themenwelten
traverse: '-4:0:0:0'
- $anchor: Postfach
shift: 86:-10
- $text: Postfach
traverse: '-3:2:0:0:0:0:0:0'
- $anchor: Computer
shift: '-15:-113'
- $text: Computer
traverse: '-9:0:0:0'
min: 5
- $size: 1
item:
type: element
container:
__$ht: element
id: 2
value: Suche nach Fragen, Diskussionen, Themen und Nutzern…
silent: true
- condition: '{{\p1}} = yes'
label: 2NNE
type: command
name: Say Command + Global Storage Management
inputs: []
silent: true
- condition: '{{\p1}} = yes'
label: PYPL
type: extract
selectorType: ai
selector:
- $matches:
- $tag: DIV
- $class: u-mls
- $class: u-hidden-tablet
- $class: u-textNoWrap
- $style: Roboto:14px:400:normal
- $content: Startseite
- $id: masthead-toggle-home
traverse: '1'
- $class: Masthead-menu
traverse: '0:1'
- $class: Masthead-banner
traverse: '1:0:1'
- $id: header-wrapper-app
traverse: '0:0:1:0:1'
- $role: banner
traverse: '0:0:1:0:1'
- $anchor: Entdecken
shift: '-112:0'
- $text: Entdecken
traverse: '-2:0:1'
- $anchor: Empfehlungen
shift: '-61:-147'
- $text: Empfehlungen
traverse: '-13:0:0:0:1:0:1'
- $anchor: Deine Startseite
shift: 186:-86
- $text: Deine Startseite
traverse: '-9:0:0:0:1:0:1'
min: 6
- $size: 1
- $parse: text
at: 0
item:
type: text
value: Startseite
container:
__$ht: element
id: 6
at: 0
param: test
default: test
silent: true
- condition: '{{\p1}} = yes'
label: FIHZ
type: export
what: all-messages
param: gpt
filename: test.txt
silent: true
- condition: '{{\p1}} = yes'
label: 6GWX
type: connection
connection: 🌐 CHAT AUTO
switchDuration: command
silent: true
- condition: '{{\p1}} = yes'
label: OSVI
type: js
args: selector
code: |-
// tip: reference arguments as args.*
const selector = args.selector || 'h1'
// tip: harpa has built-in js api
await $harpa.scroller.scroll('viewport')
await $harpa.scroller.scroll('top')
// tip: console and debugger supported
console.log('$harpa:', $harpa)
// tip: returned values are stored in params
return Array
.from(document.querySelectorAll(selector))
.map(n => n.textContent)
.join('\n')
param: test
timeout: 15000
onFailure: 6GWX
silent: true
- condition: '{{\p1}} = yes'
label: 7W96
type: control
action: frame-wide
- condition: '{{\p1}} = yes'
label: X7YI
type: prefill
message: |-
Please parse information from the web page:
{{\\\page}}
My question:
interpolate: false
- condition: '{{\p1}} = yes'
type: jump
to: 6GWX
- condition: '{{\p1}} = yes'
label: DN9X
type: group
steps:
- type: say
message: ▶️ GROUP
- condition: '{{\p1}} = yes'
label: A3Q5
type: loop
list: list
steps:
- type: say
message: |-
🔄 LOOP:
- item: {{\item}}
- index: {{\index}}
- condition: '{{\p1}} = yes'
label: XMIE
type: wait
for: idle
timeout: 15000
silent: true
- condition: '{{\p1}} = yes'
label: RFJV
type: stop
- condition: '{{\p1}} = yes'
label: W556
type: clear
silent: true
version: 1
```
STEP SELECTION GUIDELINES:
Choose from these step types based on the required functionality:
- `say`: Display a message in chat
- `ask`: Request user input
- `gpt`: Send a prompt to AI
- `navigate`: Go to a URL
- `wait`: Pause execution
- `click`: Click on page elements
- `paste`: Insert text
- `extract`: Get data from the page
- `request`: Make HTTP requests
- `js`: Run JavaScript
- `calc`: Perform calculations and set parameters
- `group`: Combine multiple steps
- `loop`: Iterate over data
- `jump`: Control execution flow
- `condition`: Execute steps conditionally
- `command`: Execute other commands
- `connection`: Switch AI providers
- `export`: Export data
FORMAT AND STYLE GUIDELINES:
- Use proper YAML indentation (2 spaces)
- Include descriptive comments for complex steps
- Use meaningful parameter names
- Structure the command in a logical sequence
- Include appropriate error handling
- Follow HARPA AI's parameter syntax with double curly braces {{parameter}}
CONSTRAINTS:
- Only include steps necessary for the requested functionality
- Ensure all references to parameters are properly formatted
- Make sure the command is complete and executable
- Do not include placeholder text that requires manual editing
- Create a self-contained command that works out of the box'
- If you use parameters, then use a \ before it, as from {param}} to
{{\param}}
Use the following information to extract ONLY useful information, but do not
adopt the format:
HARPA DOCUMENTATION:
{{array}}
Based on the MAIN QUESTION and HARPA DOCUMENTATION, please create a
complete, functional HARPA AI command in YAML format.
type: gpt
isolated: false
param: finalAnswer
dumb: false
label: FINAL ANSWER
silent: true
- type: say
message: '{{finalAnswer}}'
label: DISPLAY THE FINAL ANSWER
Notice: Please read before using
This automation command is created by a community member. HARPA AI team does not audit community commands.
Please review the command carefully and only install if you trust the creator.
Contact us
All rights reserved © HARPA AI TECHNOLOGIES LLC, 2021 — 2025
Designed and engineered in Finland 🇫🇮
CAN WE STORE COOKIES?
Our website uses cookies for the purposes of accessibility and security. They also allow us to gather statistics in order to improve the website for you. More info: Privacy Policy