⚡ Custom Deep Research Bot | v2.0
Performs in-depth online research using your custom Google queries to answer your questions. First, create up to 4 specific Google searches to gather information, then ask your actual question to get a focused AI response based on those search results
Created by Morteza H.
Updated on Jul 11, 01:50
Installed 56 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: say
message: >-
**Tip:** This command performs a thorough in-depth online search for a given
query or topic and is best used for research. It takes time and is token
hungry (~20k tokens on GPT-4o-mini per search depth).
label: TIP ABOUT COMMAND
- param: depth
message: >-
**Set the search depth.**
Search depth indicates how many search pages will be analyzed. Each depth
level generates an extra sub-query, improving search quality.
options:
- label: 🔍 SEARCH
value: 1
- label: 2 SUB-QUERIES
value: 2
- label: 3 SUB-QUERIES
value: 3
- label: 4 SUB-QUERIES
value: 4
- label: 🔗 SEARCH BY URLs
value: search_by_urls
optionsInvalid: false
type: ask
default: ''
label: SET DEPTH LEVEL
- param: all_links
message: >-
**Add Research URLs**
Paste one or more URLs you'd like the AI to focus on for this research task.
**Only these sources will be used**
You can enter multiple URLs.
options: null
optionsInvalid: false
type: ask
default: ''
label: SEARCH BY URLS
condition: '{{depth}} = search_by_urls'
- type: ask
param: placeholder
message: >-
**Model Usage Tip**
To save tokens and reduce costs, you can switch to a **cheaper AI model**
(like GPT-4.1 mini) when extracting information from the websites.
For the **final answer**, you’ll have the option to choose a **more advanced
model** to ensure the best possible results.
Click **“✅ DONE”** to proceed. You can change the model anytime via the **⋮
menu** or with **ALT + C**.
options:
- label: ✅ DONE
value: placeholder
default: ''
vision:
enabled: false
mode: area
hint: ''
send: true
optionsInvalid: false
label: TIP FOR MODEL USAGE TIP | DATA EXTRACTION
- type: jump
to: RESEARCH PROMPT INSTRUCTION
- message: >-
| Operator | Description | Example |
|----------|------------|----------|
| site: | Restricts results to a specific domain | site:arxiv.org
"transformer model" |
| -site: | Excludes a specific domain from results | AI research
-site:wikipedia.org |
| "..." | Forces exact phrase match | "machine learning architecture" |
| filetype: | Searches for specific file types | filetype:pdf "reinforcement
learning" |
| intitle: | Finds pages with specific term in title | intitle:AI benchmarks
|
| allintitle: | All terms must appear in title | allintitle:machine learning
python |
| intext: | Finds pages with specific term in body text | intext:neural
network topology |
| allintext: | All terms must appear in body text | allintext:GPT
architecture training |
| define: | Shows definition of a term | define:neural network |
| OR \| | Searches for either term | GPT-4 OR "large language model" |
| AND | Requires both terms to be present | transformer AND "attention
mechanism" |
| inurl: | Finds pages with specific term in URL |
inurl:artificial-intelligence |
| * | Wildcard for any word | "deep * model" |
| -term | Excludes a specific term | AI -chatbot |
| () | Groups search terms for complex queries | (GPT-4 OR LLaMA)
architecture |
| before: after: | Time-based filtering | AI breakthroughs after:2023 |
label: TIP ABOUT SEARCH OPERATOR
type: say
- param: userQuery1
message: >-
 Please enter {{depth}}
search queries that will be used to retrieve Google search results:
**Tip:** You have the option to improve your search query for better
results.
Choose the option below or write your own query manually:
options:
- label: ✅ IMPROVE MY QUERY
value: improve
- $custom
condition: '{{depth}} >= 1'
label: ASK FIRST SEARCH QUERY
type: ask
default: ''
optionsInvalid: false
- param: userQuery
message: >-
Great! Please describe what you're looking for in your own words.
We'll improve your query to get you the best possible results. Keep it clear
and detailed if you can.
options: null
condition: '{{userQuery1}} = improve'
label: IMRPOVE MY QUERY
type: ask
default: ''
optionsInvalid: false
- type: gpt
prompt: >-
# Prompt: Broad Coverage Single Google Search Query Generator
You are an expert Google Search Strategist. Your specialty is dissecting a
user's information need and constructing the **single broadest, yet
relevant,** Google search query using advanced syntax to maximize the number
of potential results.
OUTPUT LANGUAGE: Respond **only** in the language of the **LANGUAGE BASED ON
USER INPUT**.
TASK:
Analyze the **USER INPUT** (topic, question, or keywords) to understand the
core information need. Generate **one single** Google search query optimized
for **maximum relevant coverage and result count**.
Key Techniques & Context for Maximizing Results:
Leverage Google's advanced search operators strategically, focusing on
breadth:
- `OR`: Use extensively to include synonyms, related concepts, or
alternative phrasings for core keywords. This is the primary tool for
broadening the search.
- `( )`: Use to group `OR` statements or structure complex queries
logically, ensuring correct operator precedence. Example: `(keywordA OR
keywordB) AND (topicX OR topicY)`.
- `" "` (Single Crucial Word ONLY):** Apply quotation marks **only** if
there is **one single keyword** that is absolutely *crucial* and must appear
exactly as typed (e.g., to distinguish it from homonyms or ensure its
presence). **Do NOT use `" "` for phrases or multiple words.** If no single
word is that critical, do not use `" "`.
- Core Keywords: Identify the essential concepts but consider broader
terms or related ideas.
- `AND`: Note that Google implies `AND` between terms by default.
Explicitly writing `AND` is usually unnecessary and does not broaden the
search compared to just listing terms. Focus on `OR` for expansion.
- Other Operators (`site:`, `-`, `*`, `intitle:`, `filetype:`):** Use
these sparingly and only if they directly support the goal of broad,
relevant coverage without unduly restricting results (e.g., `site:` could be
used with `OR` across multiple relevant sites, `-` could exclude a major
*irrelevant* topic).
Inferential Capability & Query Selection:
- Keywords/Phrases: Extract core concepts. Brainstorm synonyms and related
terms suitable for `OR` combinations.
- Crucial Word Identification: Determine if *one single word* meets the
strict criteria for using `" "`. If not, omit `" "`.
- Structure: Construct the single query prioritizing broad keyword
inclusion via `OR` and logical grouping with `( )`. Balance breadth with
maintaining relevance to the user's core need.
OUTPUT FORMAT:
Present **only** the final, optimized Google search query string. Do not
include any labels, explanations, introductory/concluding remarks, or any
text other than the single query string itself.
CONSTRAINTS:
- Generate exactly **one** optimized query string designed for maximum
relevant results.
- Prioritize the use of `OR` and `( )` for breadth.
- Use `" "` **only** around a single, critical keyword if absolutely
necessary, never for phrases.
- Output **only** the query string.
- Ensure the generated query is immediately copy-paste usable in Google.
- Adhere strictly to the specified output language (based on USER INPUT).
BEGIN USER INPUT:
{{userQuery}}
label: BROAD COVERAGE SINGLE GOOGLE SEARCH QUERY GENERATOR
condition: '{{userQuery1}} = improve'
isolated: true
param: userQuery1
silent: true
- type: say
message: |-
🌐 Query created and will be used for google search:
```
{{userQuery1}}
```
label: DISPLAY IMPROVED QUERY
condition: '{{userQuery}} != —'
- param: autoQueries.boolean
message: >-
Would you like to have similar sub-queries automatically generated by an AI
based on the first sub-query you specified?
options:
- label: ✅ YES
value: 'yes'
- label: ⛔ NO
value: 'no'
vision:
mode: area
send: true
enabled: false
hint: ''
label: AUTO QUERIES BOOLEAN
condition: '{{depth}} > 1'
type: ask
default: ''
optionsInvalid: false
- type: group
steps:
- param: userQuery2
condition: '{{depth}} >= 2'
message: ''
type: ask
options: null
default: ''
- param: userQuery3
condition: '{{depth}} >= 3'
message: ''
type: ask
options: null
default: ''
- param: userQuery4
condition: '{{depth}} >= 4'
message: ''
type: ask
options: null
default: ''
label: ASK SEARCH QUERIES
condition: '{{autoQueries.boolean}} = no'
- type: ask
param: instruction
message: >-
**Set Your Research Prompt**
Enter a clear question or instruction the AI should follow when analyzing
the websites.
This prompt will be applied to **each website**, whether it comes from
**your provided URLs** or from **Google search results**.
Be as specific as possible for best results.
options: null
default: ''
vision:
enabled: false
mode: area
hint: ''
send: true
label: RESEARCH PROMPT INSTRUCTION
- param: question
message: >-
Please enter your specific query now. This query will form the basis of the
**FINAL** AI-generated detailed response. The more precisely you phrase your
question or request, the better the AI will be able to address it based on
the information it has already gathered. What would you like to know?
vision:
enabled: false
mode: area
hint: ''
send: true
label: THE REAL AND SPECIFIC QUERY FROM USER
type: ask
options: null
default: ''
- type: jump
to: EXTRACT ALL INPUT LINKS
condition: '{{depth}} = search_by_urls'
- steps:
- prompt: >-
# Prompt: Broad Coverage Multiple Google Search Queries Generator
You are an expert Google Search Strategist. Your specialty is dissecting
a user's information need and constructing **multiple broad, yet
relevant** Google search queries using advanced syntax to maximize
coverage of potential results.
OUTPUT LANGUAGE: Respond **only** in the language of the **LANGUAGE
BASED ON USER INPUT**.
TASK:
Analyze the **USER INPUT** (topic, question, or keywords) to understand
the core information need. Generate **{{depth}} different** Google
search queries optimized for **maximum relevant coverage and result
count**.
Key Techniques & Context for Maximizing Results:
Leverage Google's advanced search operators strategically, focusing on
breadth:
- `OR`: Use extensively to include synonyms, related concepts, or
alternative phrasings for core keywords. This is the primary tool for
broadening the search.
- `( )`: Use to group `OR` statements or structure complex queries
logically, ensuring correct operator precedence. Example: `(keywordA OR
keywordB) AND (topicX OR topicY)`.
- `" "` (Single Crucial Word ONLY):** Apply quotation marks **only**
if there is **one single keyword** that is absolutely *crucial* and must
appear exactly as typed (e.g., to distinguish it from homonyms or ensure
its presence). **Do NOT use `" "` for phrases or multiple words.** If no
single word is that critical, do not use `" "`.
- Core Keywords: Identify the essential concepts but consider broader
terms or related ideas.
- `AND`: Note that Google implies `AND` between terms by default.
Explicitly writing `AND` is usually unnecessary and does not broaden the
search compared to just listing terms. Focus on `OR` for expansion.
- Other Operators (`site:`, `-`, `*`, `intitle:`, `filetype:`):** Use
these sparingly and only if they directly support the goal of broad,
relevant coverage without unduly restricting results (e.g., `site:`
could be used with `OR` across multiple relevant sites, `-` could
exclude a major *irrelevant* topic).
Inferential Capability & Query Selection:
- Keywords/Phrases: Extract core concepts. Brainstorm synonyms and
related terms suitable for `OR` combinations.
- Crucial Word Identification: Determine if *one single word* meets
the strict criteria for using `" "`. If not, omit `" "`.
- Structure: Construct multiple queries prioritizing broad keyword
inclusion via `OR` and logical grouping with `( )`. Balance breadth with
maintaining relevance to the user's core need.
OUTPUT FORMAT:
Present your response as a JSON array with {{depth}} objects, each
containing a "query" field with an optimized search query. Do not
include any labels, explanations, introductory/concluding remarks, or
any text other than the JSON array.
Example JSON response:
[
{
"query": "(eco-friendly OR sustainable OR green OR biodegradable) (packaging OR containers OR wrapping) (alternatives OR options OR solutions) (small business OR SME OR startup OR entrepreneur)"
},
{
"query": "(cardboard OR paper OR bamboo OR hemp OR mushroom OR seaweed) packaging (small business OR retail OR SME) (sustainable OR eco-friendly)"
},
{
"query": "(affordable OR budget OR cost-effective) (eco-friendly OR sustainable) packaging (implementation OR transition OR switch) (small business OR SME OR startup)"
},
{
"query": "(suppliers OR vendors OR manufacturers OR brands) (eco-friendly OR green OR sustainable) packaging (small business OR SME OR local business)"
}
]
CONSTRAINTS:
- Generate exactly {{depth}} optimized query strings designed for
maximum relevant results.
- Each query should focus on a different aspect or approach to the
user's question.
- Prioritize the use of `OR` and `( )` for breadth.
- Use `" "` **only** around single, critical keywords if absolutely
necessary, never for phrases.
- Output **only** the JSON array containing the queries.
- Ensure the generated queries are immediately copy-paste usable in
Google.
- Adhere strictly to the specified output language (based on USER
INPUT).
BEGIN USER INPUT:
{{userQuery1}}
param: subqueries
condition: '{{autoQueries.boolean}} = yes'
type: gpt
label: CREATE SUBQUERIES
silent: true
dumb: false
isolated: true
- type: js
args: userQuery1, userQuery2, userQuery3, userQuery4
code: |-
function convertToJsonArray(args) {
// Extract query parameters
const queries = [
args.userQuery1 || "",
args.userQuery2 || "",
args.userQuery3 || "",
args.userQuery4 || ""
];
// Filter out queries that are too short (less than 5 characters)
const validQueries = queries
.filter(query => query && query.trim().length >= 5)
.map(query => ({ query: query.trim() }));
// The depth is automatically determined by the number of valid queries
// No need to slice the array as we only include valid queries
// Return the JavaScript array object directly
return validQueries;
}
// Execute with args from HARPA
return convertToJsonArray(args);
param: subqueries
timeout: 15000
onFailure: ''
silent: true
label: CONVERT USER QUERIES TO JSON
condition:
- '{{autoQueries.boolean}} = no'
- '{{autoQueries.boolean}} = —'
- type: calc
func: extract-json
to: subqueries
param: subqueries
index: ''
- message: >-
🧩 To give a comprehensive response, your question was split into
**{{subqueries.length}}** sub-queries.
type: say
- steps:
- value: '{{item.query}}'
func: set
param: query
format: text
type: calc
- message: '⏳ Scanning information for sub-query: "**{{query}}**".'
type: say
- param: information
value: '{{serp {{query}}}}'
type: calc
func: set
format: ''
- func: serp.extract-links
to: links
type: calc
from: information
- param: total
value: '{{links.length}}'
format: number
type: calc
func: set
- type: js
args: all_links
code: |-
function extractAllLinksAsObjects(args) {
const text = args.all_links;
if (!text) {
return [];
}
// Kombiniertes Regex Pattern für:
// 1. Markdown Links: [text](url)
// 2. Normale URLs: http(s)://example.com
const urlPattern = /$$[^$$]+$$$$([^)]+)$$|https?:\/\/[^\s<>)"]+/g;
// Array to store all link objects
const links = [];
// Set to keep track of found URLs to avoid duplicates
const foundUrls = new Set();
// Find all matches
let match;
while ((match = urlPattern.exec(text)) !== null) {
// If it's a markdown link (has capture group), use the URL from group
// Otherwise use the full match (for plain URLs)
const url = match[1] || match[0];
// Avoid duplicates
if (!foundUrls.has(url)) {
foundUrls.add(url);
links.push({ url: url });
}
}
return links;
}
return extractAllLinksAsObjects(args);
param: links
timeout: 15000
onFailure: ''
silent: true
label: EXTRACT ALL INPUT LINKS
condition: '{{depth}} = search_by_urls'
- type: calc
func: set
param: total
format: ''
value: '{{links.length}}'
label: SET LENGTH OF ARRAY LINKS
condition: '{{depth}} = search_by_urls'
- type: say
message: 🤖 Now preparing a comprehensive answer to your question.
label: SAY PREPARING MESSAGES
condition: '{{depth}} = search_by_urls'
- steps:
- value: '{{page {{item.url}}}}'
param: content
type: calc
func: set
format: ''
- type: js
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
onFailure: SAY STATUS
label: COUNT
args: content
silent: true
- param: pageUrl
value: '{{item.url}}'
type: calc
func: set
format: text
- code: |-
const regex = /^(?:https?:\/\/)?(?:www\.)?([^\/]+)/;
const testString = pageUrl;
const matches = testString.match(regex);
if (matches) {
const hostname = matches[1];
return hostname;
}
param: hostname
label: HOSTNAME
type: js
args: pageUrl
timeout: 15000
silent: true
- func: increment
param: index
type: calc
delta: 1
- 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
label: PERCENTAGE INDEX TOTAL
type: js
timeout: 15000
silent: true
- steps:
- steps:
- prompt: >-
# Enhanced Web Content Extractor
You are a precise information extraction specialist.
OUTPUT LANGUAGE: LANGUAGE BASED ON WEB PAGE CONTENT
MAIN QUESTION:
{{instruction}}
TASK: Imagine you are a Google Search engine, collecting
relevant information from various websites to form an
answer to my QUESTION.
EXTRACTION INSTRUCTIONS:
- Focus on providing detailed factual information,
statistics, and specific examples when available
- Gather all information that will be useful in
answering my question, including context, numbers, dates
and concrete details
- Extract as much useful information as possible from
the WEB PAGE CONTENT
- Avoid general phrases
- When referring to someone's opinion about something,
mention the author. Quote if appropriate
- Include supporting details and contextual information
- Capture all numerical data and specific examples
FORMAT: Respond ONLY with this JSON structure:
{
"info": "Detailed extracted information that answers the question. Include all relevant facts, figures, quotes (with attribution), dates, and statistics. Be comprehensive but focused on the question."
}
CONSTRAINTS:
- If you haven't found anything useful, don't make up
information or guess - simply say "No relevant
information found".
- 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
- Keep the style of the WEB PAGE CONTENT
Analyze the following WEB PAGE CONTENT to extract all
information relevant to MAIN QUESTION:
WEB PAGE CONTENT:
{{content}}
param: data
type: gpt
isolated: true
silent: true
label: ENHANCED WEB CONTENT EXTRACTOR
- index: first
type: calc
func: extract-json
to: data
param: data
- param: data.url
format: auto
type: calc
func: set
value: '{{item.url}}'
- param: data.title
value: '{{item.title}}'
type: calc
func: set
format: auto
- list: array
func: list-add
index: last
type: calc
item: data
- type: jump
to: SAY STATUS
condition: '{{content}} =~ ^[\s\S]{1000,}$'
label: FETCHED
type: group
label: IS IT REDDIT?
condition: >-
{{hostname}} =~
^(?!.*(reddit\.com|redd\.it)).*(?:https?:\/\/|www\.)?[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}.*$
type: group
- steps:
- 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
onFailure: ABORT NAVIGATE
label: NAVIGATE
type: js
timeout: 15000
silent: true
- type: wait
for: idle
timeout: 6000
silent: true
- type: wait
for: 2s
silent: true
label: ABORT NAVIGATE
- steps:
- param: setThread
value: '{{thread}}'
label: SET CONTENT
type: calc
func: set
format: ''
- code: |-
// For setThread
const text = args['setThread'];
const tokenCount = Math.ceil(text.length / 4);
return {
chars: text.length,
estimatedTokens: tokenCount,
estimatedWords: Math.ceil(tokenCount * 0.75)
};
onFailure: SOCIAL-MEDIA PROMPT
type: js
args: setThread
param: count
timeout: 15000
silent: true
label: COUNT
- prompt: >-
# Enhanced Web Content Extractor
You are a precise information extraction specialist.
OUTPUT LANGUAGE: LANGUAGE BASED ON WEB PAGE CONTENT
MAIN QUESTION:
{{instruction}}
TASK: Imagine you are a Google Search engine, collecting
relevant information from various websites to form an
answer to my QUESTION.
EXTRACTION INSTRUCTIONS:
- Focus on providing detailed factual information,
statistics, and specific examples when available
- Gather all information that will be useful in
answering my question, including context, numbers, dates
and concrete details
- Extract as much useful information as possible from
the WEB PAGE CONTENT
- Avoid general phrases
- When referring to someone's opinion about something,
mention the author. Quote if appropriate
- Include supporting details and contextual information
- Capture all numerical data and specific examples
FORMAT: Respond ONLY with this JSON structure:
{
"info": "Detailed extracted information that answers the question. Include all relevant facts, figures, quotes (with attribution), dates, and statistics. Be comprehensive but focused on the question."
}
CONSTRAINTS:
- If you haven't found anything useful, don't make up
information or guess - simply say "No relevant
information found".
- 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
- Keep the style of the WEB PAGE CONTENT
Analyze the following WEB PAGE CONTENT to extract all
information relevant to MAIN QUESTION:
WEB PAGE CONTENT:
{{setThread}}
type: gpt
isolated: true
param: data
silent: true
label: SOCIAL-MEDIA PROMPT
label: SOCIAL-MEDIA
condition: >-
{{hostname}} =~
^(?:reddit|facebook|twitter|telegram|discord|whatsapp)\.com$
type: group
- steps:
- param: setPageThread
value: '{{page}} {{thread}}'
type: calc
func: set
format: ''
label: SET CONTENT
- code: |-
// For setPageThread
const text = args['setPageThread'];
const tokenCount = Math.ceil(text.length / 4);
return {
chars: text.length,
estimatedTokens: tokenCount,
estimatedWords: Math.ceil(tokenCount * 0.75)
};
onFailure: NOT-SOCIAL-MEDIA PROMPT
type: js
args: setPageThread
param: count
timeout: 15000
silent: true
label: COUNT
- prompt: >-
# Enhanced Web Content Extractor
You are a precise information extraction specialist.
OUTPUT LANGUAGE: LANGUAGE BASED ON WEB PAGE CONTENT
MAIN QUESTION:
{{instruction}}
TASK: Imagine you are a Google Search engine, collecting
relevant information from various websites to form an
answer to my QUESTION.
EXTRACTION INSTRUCTIONS:
- Focus on providing detailed factual information,
statistics, and specific examples when available
- Gather all information that will be useful in
answering my question, including context, numbers, dates
and concrete details
- Extract as much useful information as possible from
the WEB PAGE CONTENT
- Avoid general phrases
- When referring to someone's opinion about something,
mention the author. Quote if appropriate
- Include supporting details and contextual information
- Capture all numerical data and specific examples
FORMAT: Respond ONLY with this JSON structure:
{
"info": "Detailed extracted information that answers the question. Include all relevant facts, figures, quotes (with attribution), dates, and statistics. Be comprehensive but focused on the question."
}
CONSTRAINTS:
- If you haven't found anything useful, don't make up
information or guess - simply say "No relevant
information found".
- 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
- Keep the style of the WEB PAGE CONTENT
Analyze the following WEB PAGE CONTENT to extract all
information relevant to MAIN QUESTION:
WEB PAGE CONTENT:
{{setPageThread}}
type: gpt
isolated: true
param: data
silent: true
label: NOT-SOCIAL-MEDIA PROMPT
condition: >-
{{hostname}} =~
^(?!(?:reddit|facebook|twitter|telegram|discord|whatsapp)\.com)[a-zA-Z0-9-]+\.[a-zA-Z]+$
label: NOT-SOCIAL-MEDIA
type: group
- 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
label: NOT FETCHED
type: group
- 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}}
condition: '{{index}} != {{total}}'
type: say
label: SAY STATUS
- 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
label: ARRAY LENGTH
type: js
args: array
timeout: 15000
onFailure: SAY STATUS
silent: true
- type: calc
func: set
param: query
format: ''
value: '{{question}}'
condition: '{{depth}} = search_by_urls'
label: SCANNED FOR QUERY
- message: |-
✅ **{{percentage}}** pages scanned for "**{{query}}**".
Last checked page: [{{hostname}}]({{pageUrl}})
- Chars: {{count.chars}}
- Estimated Tokens: {{count.estimatedTokens}}
- Estimated Words: {{count.estimatedWords}}
- Array Token length: **{{arrayLength.estimatedTokens}}**
condition: '{{index}} = {{total}}'
type: say
label: SAY STATUS
type: loop
list: links
type: loop
list: subqueries
- type: calc
func: set
param: subqueries.length
format: ''
value: '0'
condition: '{{depth}} = search_by_urls'
label: SET LENGTH TO 0 IF USING BY URLS
- message: >-
🤖 Analyzed **{{array.length}}** pages from **{{subqueries.length}}**
Google searches.
Now preparing a comprehensive answer to your question.
type: say
label: DEPTH >= 1
type: group
condition: '{{depth}} >= 1'
- type: ask
param: placeholder
message: >-
**Final Step**
This is the final response stage.
For the **best possible answer**, consider switching to a **high-quality
model** (like Gemini 2.5 Pro etc.) via the **⋮ menu** or with **ALT + C**
before continuing.
options:
- label: ✅ DONE
value: placeholder
default: ''
vision:
enabled: false
mode: area
hint: ''
send: true
optionsInvalid: false
label: TIP FOR MODEL USAGE TIP | FINAL RESPONSE
- prompt: >-
# Research Agent Prompt
You are a Research AI Agent specialized in analyzing web search results to
provide comprehensive answers.
OUTPUT LANGUAGE: {{language}}
TASK: Analyze the provided web search results (`INFORMATION FROM THE WEB`)
to generate a detailed answer to the user's question (`MY QUESTION`).
CONTEXT: The goal is to synthesize information from the provided web search
results. Extract key details, integrate findings from multiple sources, and
cite them accurately using the specified format.
OUTPUT FORMAT:
Structure your response precisely as follows:
## Key takeaway:
[Provide the single most important takeaway from the web results in one
paragraph.]
## Detailed answer:
[Present a comprehensive analysis answering the question. Use bullet points
for clarity. Integrate information from multiple sources. Cite every piece
of information using fully rendered markdown links like [➊](URL), [➋](URL),
etc., corresponding to the provided web results, placed directly within or
at the end of relevant bullet points.]
CONSTRAINTS:
- Base the answer *only* on the `INFORMATION FROM THE WEB` provided in
`INFORMATION FROM THE WEB`.
- Extract and synthesize as much relevant information as possible from the
web search results.
- Integrate insights from multiple web search sources.
- Ensure inclusion of relevant links in a fully rendered markdown format:
`[SYMBOL](URL)` (e.g., [➊](URL), [➋](URL)) to each bullet point where the
source is used.
- Each unique URL must correspond to only one symbol: ➊ ➋ ➌ ➍ ➎ ➏ ➐ ➑ ➒ ➓.
- If multiple sources repeat the same information, describe it and cite
all applicable sources together (e.g., [➊](URL)[➌](URL)[➎](URL)).
- Do *not* hallucinate facts or information.
- Do not use any tools other than analyzing the provided `INFORMATION FROM
THE WEB`.
- Avoid general phrases; be specific and detailed.
- Adhere strictly to the "Key takeaway" and "Detailed answer" structure.
MY QUESTION:
{{question}}
INFORMATION FROM THE WEB:
{{array}}
COMPREHENSIVE RESPONSE WITH SOURCE LINKS ([➊](URL)):
label: FINAL ANSWER
type: gpt
isolated: true
param: ''
dumb: false
- type: js
args: array
code: |-
function formatMarkdownLinks(args) {
// Hole die Daten aus den Argumenten
let data = args.array;
// Wenn data ein String ist und nach JSON aussieht, versuche es zu parsen
if (typeof data === 'string' && (data.trim().startsWith('[') || data.trim().startsWith('{'))) {
try {
data = JSON.parse(data);
} catch (e) {
// Kein gültiges JSON, mit Textverarbeitung fortfahren
}
}
// Prüfe, ob wir ein Array von Objekten mit URL-Eigenschaft haben
if (Array.isArray(data) && data.length > 0 && typeof data[0] === 'object' && data[0].url) {
// Kreiszahlen für die Formatierung
const circleNumbers = "➊➋➌➍➎➏➐➑➒➓⓫⓬⓭⓮⓯⓰⓱⓲⓳⓴㉑㉒㉓㉔㉕㉖㉗㉘㉙㉚㉛㉜㉝㉞㉟㊱㊲㊳㊴㊵".split("");
// Array für eindeutige Links
const uniqueLinks = [];
// Extrahiere URLs aus jedem Objekt im Array
data.forEach((item, index) => {
if (item.url && index < circleNumbers.length) {
try {
const url = item.url;
const urlObj = new URL(url);
const fullPath = urlObj.hostname.replace('www.', '') + urlObj.pathname;
const shortPath = fullPath.length > 42
? fullPath.slice(0, 42) + '...'
: fullPath.replace(/\/$/, '');
// Füge den formatierten Link zum uniqueLinks-Array hinzu
uniqueLinks.push(`[${circleNumbers[index]}](${url}) [${shortPath}](${url})`);
} catch (e) {
// Ungültige URLs überspringen
}
}
});
// Gib die formatierten Links zurück oder eine Nachricht, wenn keine gefunden wurden
return uniqueLinks.length > 0
? uniqueLinks.join('\n')
: "Keine URLs in den Daten gefunden.";
}
// Wenn es sich nicht um ein JSON-Array von Objekten mit URLs handelt,
// verwende die ursprüngliche Funktion
const text = typeof data === 'string' ? data : JSON.stringify(data);
if (!text) return "Kein Text zum Extrahieren von Links bereitgestellt.";
// Regex pattern für Links mit ➊, ➋, ➌ etc.
const linkPattern = /\[(➊|➋|➌|➍|➎|➏|➐|➑|➒|➓|⓫|⓬|⓭|⓮|⓯|⓰|⓱|⓲|⓳|⓴|㉑|㉒|㉓|㉔|㉕|㉖|㉗|㉘|㉙|㉚|㉛|㉜|㉝|㉞|㉟|㊱|㊲|㊳|㊴|㊵)\]\(([^)]+)\)/g;
// Erweiterte Liste von Kreiszahlen
const circleNumbers = "➊➋➌➍➎➏➐➑➒➓⓫⓬⓭⓮⓯⓰⓱⓲⓳⓴㉑㉒㉓㉔㉕㉖㉗㉘㉙㉚㉛㉜㉝㉞㉟㊱㊲㊳㊴㊵".split("");
// Map zur Speicherung eindeutiger Links
const uniqueLinks = new Map();
// Alle Übereinstimmungen finden
let match;
let currentIndex = 0;
while ((match = linkPattern.exec(text)) !== null) {
try {
const symbol = match[1];
const url = match[2];
// URL-Objekt erstellen zur Validierung und Analyse
const urlObj = new URL(url);
// Domain und Pfad extrahieren (ohne www.)
const fullPath = urlObj.hostname.replace('www.', '') + urlObj.pathname;
// Kürzen, wenn länger als 42 Zeichen
const shortPath = fullPath.length > 42
? fullPath.slice(0, 42) + '...'
: fullPath.replace(/\/$/, '');
// Erstelle einen Key, der Symbol und URL kombiniert
const key = `${symbol}:${url}`;
// Jede Übereinstimmung wird mit einem eindeutigen Symbol und URL-Kombination hinzugefügt
if (!uniqueLinks.has(key)) {
uniqueLinks.set(key, `[${symbol}](${url}) [${shortPath}](${url})`);
currentIndex++;
}
} catch (e) {
// Ungültige URLs überspringen
}
}
// Wenn keine formatierten Links gefunden wurden, nach einfachen URLs suchen
if (uniqueLinks.size === 0) {
const urlPattern = /(https?:\/\/[^\s\)\]\'",:;]+)/g;
let urlMatch;
currentIndex = 0;
while ((urlMatch = urlPattern.exec(text)) !== null) {
try {
const url = urlMatch[1].replace(/[.,;:?!]+$/, ''); // URL bereinigen
// URL-Objekt erstellen zur Validierung und Analyse
const urlObj = new URL(url);
// Domain und Pfad extrahieren (ohne www.)
const fullPath = urlObj.hostname.replace('www.', '') + urlObj.pathname;
// Kürzen, wenn länger als 42 Zeichen
const shortPath = fullPath.length > 42
? fullPath.slice(0, 42) + '...'
: fullPath.replace(/\/$/, '');
// Stoppe, wenn wir mehr URLs haben als Kreiszahlen
if (currentIndex >= circleNumbers.length) break;
// Jede URL wird mit einem eigenen Symbol hinzugefügt
const key = `${currentIndex}:${url}`;
uniqueLinks.set(key, `[${circleNumbers[currentIndex]}](${url}) [${shortPath}](${url})`);
currentIndex++;
} catch (e) {
// Ungültige URLs überspringen
}
}
}
// Formatierte Ausgabe zurückgeben oder eine Nachricht, wenn keine Links gefunden wurden
return uniqueLinks.size > 0
? Array.from(uniqueLinks.values()).join('\n')
: "Keine nummerierten Referenzlinks im Text gefunden.";
}
return formatMarkdownLinks(args);
param: sources
timeout: 15000
onFailure: DISPLAY SOURCES
label: EXTRACT SOURCES
silent: true
- type: say
message: |-
 Sources
{{sources}}
label: DISPLAY SOURCES
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