filter-rental-search

rent.591.com.tw/filter-rental-search-fq54br · updated May 21, 2026

MDX-style export adds YAML metadata + attribution linking explainx.ai and this canonical listing URL.

$browse install rent.591.com.tw/filter-rental-search-fq54br
0 commentsdiscussion
summary

Apply user-supplied filters (region, district, price, room count, property type, area, amenities, sort) on rent.591.com.tw and return matching rental listings with title, price, layout, area, floor, address, MRT distance, and detail URL.

skill.md
name
filter-rental-search
title
591租屋網 Filter Rental Search
description
>- Apply user-supplied filters (region, district, price, room count, property type, area, amenities, sort) on rent.591.com.tw and return matching rental listings with title, price, layout, area, floor, address, MRT distance, and detail URL.
website
rent.591.com.tw
category
real-estate
tags
- real-estate - rental - taiwan - '591' - search - filters - read-only
source
'browserbase: agent-runtime 2026-05-20'
updated
'2026-05-20'
recommended_method
url-param
alternative_methods
- method: browser rationale: >- The filter UI is fully clickable via .item-info card extraction + label-text matching, but every UI click maps 1:1 to a URL query-string param — constructing the URL directly skips 5-10 click-then-snapshot turns. Use the click path only to discover an unknown filter slug, then prefer URL params for production. - method: api rationale: >- Direct API path /v3/web/rent/list observed in the SSR X-Server-Monitor header is 404 to external clients — only callable server-side from the Nuxt SSR layer. No public JSON endpoint exists for filtered listings; the SSR'd HTML is the canonical data surface.
verified
false
proxies
false

591租屋網 — Filter Rental Search

Purpose

Apply user-supplied rental filters (city, district, price range, room count, property type, area, amenities, sort order) on rent.591.com.tw (Taiwan's largest rental listings site) and return the matching listings — title, price, layout, area (坪 / ping), floor, address, district, distance to nearest metro station, agent/owner, last-updated time, and the canonical detail-page URL. Read-only — never contacts landlords, never logs in.

When to Use

  • "Find me 2-bedroom apartments in Taipei 大安區 / 信義區 between NT$15,000 and NT$25,000."
  • "Cheapest studios (套房) near a metro station in 新北市 板橋區."
  • Rental-market monitoring: hourly/daily polling for new listings (sort=posttime_desc&other=newPost).
  • Cross-region rent comparison (region 1 = Taipei, 3 = New Taipei, 8 = Taichung, 17 = Kaohsiung).
  • Anywhere you would otherwise scrape 591 with brittle CSS selectors — filter via URL params, then read .item-info cards.

Workflow

The recommended path is URL-param filtering + DOM extraction — every filter the site UI exposes is also a query-string param, so a single browse open with the right URL replaces 5–10 click-then-snapshot turns. The page is server-rendered (Nuxt SSR) so .item-info cards are present in the initial DOM before any JS runs. Anti-bot is minimal — bare Browserbase sessions (no --proxies, no --verified) successfully load and render listings. Ship stealth/proxies only if you start seeing ratelimit pages.

1. Construct the filter URL

Base: https://rent.591.com.tw/list?{params}

ParamMeaningSyntaxExample
regionCity (required, single value)integerregion=1 (台北市)
sectionDistrict(s) within regioncomma-list of integerssection=5,7 (大安+信義)
kindRental typecomma-list of integerskind=2,1 (studio + whole apt)
priceRent range NT$/monthMIN_MAX; comma-join multiple rangesprice=15000_25000 or price=0_5000,30000_40000
layoutRoom countcomma-list of 14layout=2,3 (2房 or 3房)
shapeBuilding shapecomma-list of integersshape=1,3,4 (公寓+透天+別墅)
acreageFloor area (坪 / ping)MIN_MAXacreage=10_30
otherAmenity / status flagscomma-list of slug strings (see enum below)other=near_subway,pet,lift
sortResult orderingslugsort=posttime_desc (newest first; default = "預設排序" = no sort param)

region enum (Taiwan-wide; gaps 9, 16, 18, 20 are non-Taiwan codes that render an empty title):

IDRegionIDRegionIDRegion
1台北市 (Taipei City)2基隆市 (Keelung)3新北市 (New Taipei)
4新竹市 (Hsinchu City)5新竹縣 (Hsinchu County)6桃園市 (Taoyuan)
7苗栗縣 (Miaoli)8台中市 (Taichung)10彰化縣 (Changhua)
11南投縣 (Nantou)12嘉義市 (Chiayi City)13嘉義縣 (Chiayi County)
14雲林縣 (Yunlin)15台南市 (Tainan)17高雄市 (Kaohsiung)
19屏東縣 (Pingtung)21宜蘭縣 (Yilan)22台東縣 (Taitung)
23花蓮縣 (Hualien)24澎湖縣 (Penghu)25金門縣 (Kinmen)

kind enum (only 14, 810 are valid for rent.591.com.tw; 5/6/7/11 redirect to other 591 properties — store.591, office.591, factory listings, sale-only land — and lose the rent filter UI):

IDType
0不限 — all rental types (default; equivalent to omitting kind)
1整層住家 — whole apartment
2獨立套房 — independent studio (own bath/kitchen)
3分租套房 — sublet studio (own bath, shared common area)
4雅房 — private bedroom with shared bath
8車位 — parking space
9住宅 — residential aggregate
10套房 — studio aggregate (covers 2 + 3)

section enum (region-specific; always re-derive per region — IDs are not stable across regions and you cannot port a Taipei section=5 to a New Taipei query). For 台北市 (region 1), discovered:

IDDistrictIDDistrictIDDistrict
1中正區2大同區3中山區
4松山區5大安區6萬華區
7信義區8士林區9北投區
10內湖區11南港區12文山區

For other regions, derive the mapping by clicking each district checkbox once and reading the URL — see the click-and-read snippet in Browser fallback.

other enum (amenity / status flags, comma-joined):

SlugLabel
newPost新上架 (newly listed)
near_subway近捷運 (near MRT/metro)
pet可養寵物 (pet-friendly)
cook可開伙 (cooking allowed)
cartplace有車位 (parking included)
lift有電梯 (elevator)
balcony_1有陽台 (balcony)
lease可短期租賃 (short-term OK)
social-housing社會住宅 (social housing)
rental-subsidy租金補貼 (subsidy-eligible)
elderly-friendly高齡友善 (elderly-friendly)
tax-deductible可報稅 (tax-deductible)
naturalization可入籍 (household-registration eligible)

shape enum (partial — derived by multi-select reverse-mapping shape=1,3,4 → 公寓 + 透天厝 + 別墅):

IDType
1公寓 (walk-up apartment)
3透天厝 (townhouse / single-family)
4別墅 (villa)

(2 and 5+ exist but were not exhaustively probed; click-and-read the UI if you need them.)

2. Open the URL

SID=$(browse cloud sessions create --keep-alive | node -e "let s='';process.stdin.on('data',c=>s+=c).on('end',()=>process.stdout.write(JSON.parse(s).id))")
export BROWSE_SESSION="$SID"

URL="https://rent.591.com.tw/list?region=1&kind=2,1&price=15000_25000&layout=2&sort=posttime_desc"
browse open "$URL" --remote
browse wait load --remote
# Listings are SSR'd into the initial HTML; no extra wait usually needed.
# Add `browse wait timeout 2000` only if the popup tooltip is blocking your evals.

The site auto-rewrites legacy param names (multiPriceprice, multiRoomlayout, etc.) — don't rely on the auto-rewrite, just send the canonical names from the table above.

3. Extract the total count

browse eval --remote 'JSON.stringify({total: (document.body.innerText.match(/已為你找到\s*([0-9,]+)\s*間/) || [])[1]});'
# → { "total": "1,525" }

The total reflects the full filtered match across all pages (not just the 30 cards rendered on page 1).

4. Extract the listing cards

The first page of the SSR'd response renders 30 listing cards in .item-info. Each card has a child <a> whose href is the canonical detail URL.

browse eval --remote '
const total = (document.body.innerText.match(/已為你找到\s*([0-9,]+)\s*間/) || [])[1];
const cards = [...document.querySelectorAll(".item-info")];
const listings = cards.map(c => {
  const lines = (c.innerText || "").split("\n").map(l=>l.trim()).filter(Boolean);
  const url = c.href || c.querySelector("a")?.href;
  const id = url?.match(/\/(\d{6,})$/)?.[1];
  const priceMatch = c.innerText.match(/([0-9,]+)\s*元\/月/);
  // Layout/area/floor: a single line like "獨立套房14.9坪6F/7F" or "整層住家2房1廳1衛20坪3F/5F"
  const sizeLine = lines.find(l => /\d+(\.\d+)?坪/.test(l)) || "";
  const ping = parseFloat((sizeLine.match(/(\d+(\.\d+)?)坪/) || [])[1]) || null;
  const floor = (sizeLine.match(/([0-9頂樓加蓋]+F\/\d+F)/) || [])[1] || null;
  const layoutStr = (sizeLine.match(/(\d+房\d*廳\d*衛|獨立套房|分租套房|雅房|整層住家|開放式|樓中樓)/) || [])[1] || null;
  // Address: the line containing "區-" (district-street)
  const address = lines.find(l => /區-/.test(l)) || null;
  // Distance to MRT: line starting with "距"
  const distance = lines.find(l => l.startsWith("距")) || null;
  return { id, url, title: lines[0], layout: layoutStr, area_ping: ping, floor, address, distance, price_ntd: priceMatch ? +priceMatch[1].replace(/,/g, "") : null };
});
JSON.stringify({total, count: listings.length, listings});
'

The first line of c.innerText is the listing title; the last line containing the substring 元/月 is the price; one line near the bottom of the card has the agent/owner + last-updated stamp + "昨日 X 人瀏覽" view counter (skip if not needed).

5. Pagination (if needed)

To paginate beyond the first 30 cards, append &firstRow=N with N in steps of 30: firstRow=30, firstRow=60, etc. (Some 591 mirrors call this pg= or page=, but firstRow is the canonical for rent.591.) The total count from step 3 tells you how many pages to walk.

6. Release the session

browse cloud sessions update "$SID" --status REQUEST_RELEASE

Browser fallback (when you need a filter slug you don't have)

If the user asks for a filter you don't yet have a URL slug for (e.g., a section value in 台中市), open the bare filter page, click the checkbox, and read the URL:

browse open "https://rent.591.com.tw/list?region=8" --remote   # Taichung
sleep 2
browse eval --remote '[...document.querySelectorAll("label")].find(l=>l.innerText?.trim()==="北屯區")?.click(); "ok";'
sleep 1
browse get url --remote   # → ...&section=NN

This same recipe works for any filter chip — kind, shape, other, etc. The Vue store mutation is what builds the URL, so the canonical slug always appears in window.location.search immediately after the click.

Site-Specific Gotchas

  • browse snapshot consistently fails on this site with a Stagehand parser error (observed across 8 attempts in iter 1, 2 of testing). Use browse eval for DOM extraction directly. The accessibility-tree path is not viable; the explicit DOM-query path is.
  • No serious anti-bot. Bare Browserbase sessions (no --proxies, no --verified) successfully load filtered pages and render the full .item-info grid. The metadata for this skill reflects that — verified: false, proxies: false. Add stealth only if you observe Cloudfront 403s during sustained polling.
  • A "RoboForm 密碼管理工具" tooltip popup overlays the right side of the viewport on first page load. It does NOT block DOM evaluation, but it WILL show up in screenshots. Dismiss programmatically before screenshotting:
    [...document.querySelectorAll("*")].filter(e=>e.children.length===0 && e.innerText?.trim()==="我知道了,不再提示")[0]?.click();
    
  • Param auto-rewrite. Sending multiPrice=15000_25000&multiRoom=2&kind=2,1 rewrites to kind=2,1&price=15000_25000&layout=2 on the next render. Don't rely on legacy names — use the canonical price, layout, section, shape, acreage, other, sort listed in the table.
  • section IDs are region-scoped and NOT portable. A Taipei section=5 (大安區) is meaningless in another region. Always re-derive section IDs per region with the click-and-read fallback recipe.
  • region codes 9, 16, 18, 20 render a generic "租屋 | 房屋出租" title with no city name — these slots are reserved (probably for outlying Taiwan municipalities not listed in the rental flow). Only use the regions in the enum table; treat anything else as undefined.
  • kind 5/6/7/11 redirect to sibling 591 properties (store.591 for retail/storefront, separate office/factory/land sites) and lose the rental filter UI. The valid rental kinds are 0, 1, 2, 3, 4, 8, 9, 10.
  • First page returns 30 cards regardless of filter strictness (vs. typical 12–20 elsewhere). Page 2 starts at firstRow=30. Don't assume Craigslist-style 360-card batches.
  • Bare https://rent.591.com.tw/ sets a urlJumpIp cookie and 301-redirects to /list?region=N based on the request IP (in our trace, region=1 from any non-Taiwan IP — Taipei is the global default). Pass region= explicitly to lock the city. The redirect chain is: //list/list?region=1.
  • Sort param's "默認" (default) is the empty value — sending sort= or omitting sort both yield the site's blended-relevance ordering (which up-ranks "優選好屋" promoted listings). Use sort=posttime_desc if you want true newest-first across the whole result set.
  • Listing IDs in https://rent.591.com.tw/{id} are 7–8 digit integers, not slugs. The detail page itself lives at rent.591.com.tw/{id} with no slug, no path segments.
  • Total count is in localized text only. Use the regex /已為你找到\s*([0-9,]+)\s*間/ against document.body.innerText — there is no data-* attribute exposing the integer count cleanly. The count includes the comma thousand-separator and you must strip it before parsing.
  • Card innerText is multi-line and order-sensitive. Title is lines[0]. Size/floor/type is the line matching /\d+坪/. Address is the line matching /區-/. Don't rely on line indices — match by content shape.
  • Direct API path tried and failed. GET https://rent.591.com.tw/v3/web/rent/list?region=1 (the path observed in the Cloudfront X-Server-Monitor header) returns a 404 to external clients — it's only callable server-side from the SSR layer. Don't waste turns trying to find a JSON endpoint; the SSR'd HTML is the canonical data surface.

Expected Output

{
  "success": true,
  "filters_applied": {
    "region": 1,
    "region_name": "台北市",
    "section": [5, 7],
    "section_names": ["大安區", "信義區"],
    "kind": [1, 2],
    "kind_names": ["整層住家", "獨立套房"],
    "price_min": 15000,
    "price_max": 25000,
    "layout": [2],
    "other": ["near_subway", "lift"],
    "sort": "posttime_desc"
  },
  "url": "https://rent.591.com.tw/list?region=1&section=5,7&kind=2,1&price=15000_25000&layout=2&other=near_subway,lift&sort=posttime_desc",
  "total_matches": 1525,
  "page_size": 30,
  "listings": [
    {
      "id": "21297268",
      "url": "https://rent.591.com.tw/21297268",
      "title": "中山站、電梯套房、台水電、獨洗、有陽台、代收垃圾",
      "layout": "獨立套房",
      "area_ping": 14.9,
      "floor": "6F/7F",
      "address": "中山區-林森北路259巷",
      "distance": "距中山459公尺",
      "price_ntd": 23500
    }
  ],
  "error_reasoning": null
}

Outcome shapes:

// Zero matches (filter combination too restrictive)
{ "success": true, "total_matches": 0, "listings": [], "url": "...", "filters_applied": {...} }

// kind redirected to a sibling property (e.g. user passed kind=5 / store)
{ "success": false, "reason": "kind_not_supported_by_rent_591", "redirected_to": "store.591.com.tw" }

// region code in the reserved-but-unused slots (9/16/18/20)
{ "success": false, "reason": "invalid_region", "region": 16 }
how to use filter-rental-search

How to use filter-rental-search on Cursor

AI-first code editor with Composer

1

Prerequisites

Before installing skills in Cursor, ensure your development environment meets these requirements:

  • Cursor installed and configured on your development machine
  • Node.js version 16.0+ with npm package manager (verify with node --version)
  • Active project directory or workspace where you want to add filter-rental-search
2

Execute installation command

Execute the skills CLI command in your project's root directory to begin installation:

$browse install rent.591.com.tw/filter-rental-search-fq54br

The skills CLI fetches filter-rental-search from GitHub repository rent.591.com.tw/filter-rental-search-fq54br and configures it for Cursor.

3

Select Cursor when prompted

The CLI will show a list of available agents. Use arrow keys to navigate and space to select Cursor:

◆ Which agents do you want to install to?
│ ── Universal (.agents/skills) ── always included ────
│ • Amp
│ • Antigravity
│ • Cline
│ • Codex
│ ●Cursor(selected)
│ • Cursor
│ • Windsurf
4

Verify installation

Confirm successful installation by checking the skill directory location:

.cursor/skills/filter-rental-search

Reload or restart Cursor to activate filter-rental-search. Access the skill through slash commands (e.g., /filter-rental-search) or your agent's skill management interface.

Security & Verification Notice

We perform automated surface-level scans (Gen AI Scanner, Socket, Snyk) during installation. These checks detect common vulnerabilities but do not guarantee complete security. Always review skill source code and verify the publisher's reputation before production use.

Skills execute code in your development environment. Always verify the publisher's identity, review recent commits, and test in isolated environments before production deployment.

List & Monetize Your Skill

Submit your Claude Code skill and start earning

GET_STARTED →

Use Cases

Task Automation & Efficiency

Automate repetitive workflows and reduce manual effort

Example

Generate reports, summarize documents, draft communications

Save 3-5 hours per week on routine tasks

Knowledge Enhancement

Learn new skills, understand complex topics, get expert guidance

Example

Explain concepts, provide examples, suggest learning resources

Accelerate learning and skill development by 2x

Quality Improvement

Enhance output quality through reviews, suggestions, and refinements

Example

Review drafts, suggest improvements, catch errors

Improve work quality by 30-40% with less effort

Implementation Guide

Prerequisites

  • Claude Desktop or compatible AI client with skill support
  • Clear understanding of task or problem to solve
  • Willingness to iterate and refine outputs

Time Estimate

15-45 minutes depending on use case complexity

Installation Steps

  1. 1.Install skill using provided installation command
  2. 2.Test with simple use case relevant to your work
  3. 3.Evaluate output quality and relevance
  4. 4.Iterate on prompts to improve results
  5. 5.Integrate into regular workflow if valuable

Common Pitfalls

  • Expecting perfect results without iteration
  • Not providing enough context in prompts
  • Using skill for tasks outside its intended scope
  • Accepting outputs without review and validation

Best Practices

✓ Do

  • +Start with clear, specific prompts
  • +Provide relevant context and constraints
  • +Review and refine all outputs before using
  • +Iterate to improve output quality
  • +Document successful prompt patterns

✗ Don't

  • Don't use without understanding skill limitations
  • Don't skip validation of outputs
  • Don't share sensitive information in prompts
  • Don't expect skill to replace human judgment

💡 Pro Tips

  • Be specific about desired format and style
  • Ask for multiple options to choose from
  • Request explanations to understand reasoning
  • Combine AI efficiency with human expertise

When to Use This

✓ Use When

Use when skill capabilities match your task, clear ROI on time saved, and you can validate outputs. Best for repetitive tasks, learning, and quality improvement.

✗ Avoid When

Avoid when task requires deep expertise you can't validate, involves sensitive decisions, or when learning process is more valuable than speed of completion.

Learning Path

  1. 1Familiarize yourself with skill capabilities and limitations
  2. 2Start with low-risk, non-critical tasks
  3. 3Progress to more complex and valuable use cases
  4. 4Build expertise through regular use and experimentation

Discussion

Product Hunt–style comments (not star reviews)
  • No comments yet — start the thread.
general reviews

Ratings

4.664 reviews
  • Ganesh Mohane· Dec 16, 2024

    filter-rental-search has been reliable in day-to-day use. Documentation quality is above average for community skills.

  • Meera Bansal· Dec 16, 2024

    Solid pick for teams standardizing on skills: filter-rental-search is focused, and the summary matches what you get after install.

  • Advait Lopez· Dec 16, 2024

    filter-rental-search has been reliable in day-to-day use. Documentation quality is above average for community skills.

  • Amina Malhotra· Dec 12, 2024

    filter-rental-search reduced setup friction for our internal harness; good balance of opinion and flexibility.

  • Yuki Abebe· Dec 12, 2024

    I recommend filter-rental-search for anyone iterating fast on agent tooling; clear intent and a small, reviewable surface area.

  • Advait Yang· Dec 12, 2024

    Keeps context tight: filter-rental-search is the kind of skill you can hand to a new teammate without a long onboarding doc.

  • Sakshi Patil· Nov 7, 2024

    filter-rental-search reduced setup friction for our internal harness; good balance of opinion and flexibility.

  • Aanya Smith· Nov 7, 2024

    filter-rental-search reduced setup friction for our internal harness; good balance of opinion and flexibility.

  • Kofi Dixit· Nov 3, 2024

    filter-rental-search has been reliable in day-to-day use. Documentation quality is above average for community skills.

  • Aisha Srinivasan· Nov 3, 2024

    Useful defaults in filter-rental-search — fewer surprises than typical one-off scripts, and it plays nicely with `npx skills` flows.

showing 1-10 of 64

1 / 7