# Airbnb Reviews & Guest Sentiment Miner (`apt_marble/airbnb-reviews-scraper`) Actor

Collect every Airbnb review for any listing: full text, 1-5 star rating, exact date, language, reviewer and the host's reply. Each listing also gets its published rating breakdown, star histogram, topic tags, reply rate, monthly review volume and language mix. No account needed.

- **URL**: https://apify.com/apt\_marble/airbnb-reviews-scraper.md
- **Developed by:** [Hamza](https://apify.com/apt_marble) (community)
- **Categories:** Travel, Real estate, Automation
- **Stats:** 2 total users, 1 monthly users, 100.0% runs succeeded, 0 bookmarks
- **User rating**: No ratings yet

## Pricing

from $0.30 / 1,000 review saveds

This Actor is paid per event. You are not charged for the Apify platform usage, but only a fixed price for specific events.

Learn more: https://docs.apify.com/platform/actors/running/actors-in-store#pay-per-event

## What's an Apify Actor?

Actors are web data automations that power AI and operations. They run on the Apify platform to scrape websites, process data, connect APIs, and automate workflows.
In Batch mode, an Actor accepts a well-defined JSON input, performs an action which can take anything from a few seconds to a few hours,
and optionally produces a well-defined JSON output, datasets with results, or files in key-value store.
In Standby mode, an Actor provides a web server which can be used as a website, API, or an MCP server.
Actors are written with capital "A".

## How to integrate an Actor?

If asked about integration, you help developers integrate Actors into their projects.
You adapt to their stack and deliver integrations that are safe, well-documented, and production-ready.
The best way to integrate Actors is as follows.

- **AI agents and MCP clients** — the [Apify MCP server](https://docs.apify.com/integrations/mcp.md) at `https://mcp.apify.com` (remote, streamable HTTP, OAuth on first use).
- **Agentic workflows and local Actor development** — [Agent Skills](https://apify.com/.well-known/agent-skills/index.json) with the [Apify CLI](https://docs.apify.com/cli/docs.md): `npm install -g apify-cli`, then `apify login`.
- **JavaScript/TypeScript projects** — the official [JS/TS client](https://docs.apify.com/api/client/js/docs.md): `npm install apify-client`.
- **Python projects** — the official [Python client](https://docs.apify.com/api/client/python/docs.md): `pip install apify-client`.
- **Any other language** — the [REST API](https://docs.apify.com/api/v2.md).

For usage examples, see the [API](#api) section below.

For more details, see Apify documentation as [Markdown index](https://docs.apify.com/llms.txt) and [Markdown full-text](https://docs.apify.com/llms-full.txt).

# README

## Airbnb Reviews & Guest Sentiment Miner

Turn any Airbnb listing into a clean, structured review dataset. Paste a room
link and get back every guest review with its full original text, its 1–5 star
rating, the exact date it was written, the language it was written in, the
reviewer's first name and profile, and the host's public reply where there is
one. Alongside the reviews, each listing gets a single summary row carrying the
figures Airbnb publishes on the page — the overall rating, the six category
scores, the 1–5 star histogram and Airbnb's own topic tags — plus derived
measures you would otherwise build by hand: host reply rate, review volume month
by month, the rating trend over time, the share of reviews under four stars and
the language mix of the guests. No Airbnb account, no login, nothing to connect.

### What you can do with it

- **Audit a property before you buy or manage it.** Read every review a listing
  has ever received, in order, and see whether the 4.9 headline hides a run of
  cleanliness complaints in the last six months.
- **Benchmark a comp set.** Point it at ten rivals in the same market and compare
  category scores, share of sub-4-star reviews and reply rates side by side.
- **Estimate demand without guessing.** Review volume month by month is the
  closest public signal for how often a listing is actually booked, and it comes
  back as a ready-to-chart series.
- **Mine complaints for product and operations work.** Filter to reviews of three
  stars or fewer and read only what went wrong, across a whole portfolio.
- **Watch a market's guest sentiment over time.** Schedule a run with a start
  date and append only the reviews written since the last one.
- **Feed a summariser or a sentiment model.** Review text arrives complete and in
  its original language, with Airbnb's own translation attached when you ask for
  it — real writing, not truncated previews.

### What you get

Two kinds of row, told apart by `recordType`. One row per review:

```json
{
  "recordType": "review",
  "reviewId": "1626871481184794267",
  "listingId": "5398166",
  "listingUrl": "https://www.airbnb.com/rooms/5398166",
  "rating": 5,
  "reviewText": "Great little flat, five minutes from the metro and spotless.",
  "reviewLanguage": "en",
  "createdAt": "2026-02-22T15:43:13Z",
  "localizedDate": "February 2026",
  "reviewerFirstName": "Sophia",
  "reviewerId": "586517841",
  "reviewerContextualUserId": "1485235383355693118",
  "reviewerPictureUrl": "https://a0.muscache.com/im/pictures/user/User/original/e909ba2c.jpeg",
  "reviewerProfilePath": "/users/profile/1485235383355693118",
  "reviewerIsSuperhost": false,
  "reviewerTenureText": "2 years on Airbnb",
  "reviewHighlight": "Stayed a few nights",
  "highlightType": "LENGTH_OF_STAY",
  "hostResponseText": null,
  "hostResponseDateText": null,
  "hostId": "10214195",
  "hostFirstName": "Pierre",
  "reviewPhotoUrls": [],
  "scrapedAt": "2026-07-29T09:14:02.117Z"
}
```

And one summary row per listing:

```json
{
  "recordType": "listingSummary",
  "listingId": "5398166",
  "listingUrl": "https://www.airbnb.com/rooms/5398166",
  "listingTitle": "Nice apartment center of paris",
  "propertyType": "Entire rental unit",
  "personCapacity": 4,
  "bedrooms": 1,
  "beds": 2,
  "bathrooms": 1.5,
  "overallRating": 4.68,
  "overallCount": 78,
  "isGuestFavorite": false,
  "categoryRatings": [
    { "label": "Cleanliness", "rating": 4.4, "accessibilityLabel": "Rated 4.4 out of 5 stars for cleanliness" },
    { "label": "Value", "rating": 4.6, "accessibilityLabel": "Rated 4.6 out of 5 stars for value" }
  ],
  "ratingHistogram": [
    { "stars": 5, "percentage": 0.7435897435897436, "localizedRating": "74%" },
    { "stars": 4, "percentage": 0.21794871794871795, "localizedRating": "22%" }
  ],
  "topicTags": [
    { "name": "LOCATION", "localizedName": "Location", "count": 57 },
    { "name": "HOSPITALITY", "localizedName": "Hospitality", "count": 39 }
  ],
  "reviewsRetrieved": 78,
  "sampleSize": 78,
  "sampleAverageRating": 4.71,
  "publishedReviewCount": 78,
  "hostResponseRate": 0.04,
  "medianResponseLagDays": 0,
  "responseMetricsFromSampleOnly": true,
  "reviewsPerMonth": [ { "month": "2026-01", "reviews": 3 }, { "month": "2026-02", "reviews": 2 } ],
  "monthlyRatingTrend": [ { "month": "2026-02", "reviews": 2, "averageRating": 5 } ],
  "shareSub4": 0.04,
  "languageMix": [ { "language": "en", "reviews": 23, "share": 0.46 } ],
  "scrapedAt": "2026-07-29T09:14:02.117Z"
}
```

### Input reference

| Field | Type | Default | What it does |
| --- | --- | --- | --- |
| `listingUrls` | list of strings | one example listing | Room links to collect reviews for, exactly as they appear in the address bar. Up to 500 per run. |
| `listingIds` | list of strings | empty | The same listings as bare numeric ids, for when they come out of a spreadsheet. |
| `query` | string | empty | A market to collect listings from when you have not supplied any links — "Paris, France", "Austin, Texas". |
| `bbox` | object | empty | Narrows the market to a rectangle on the map, as `{"neLat": 48.9, "neLng": 2.45, "swLat": 48.8, "swLng": 2.25}`. Must be paired with a market name. |
| `maxListings` | integer | `50` | How many listings one run works through, up to 500. |
| `maxReviewsPerListing` | integer | `200` | Upper bound on reviews per listing, up to 2,000. Set it to `0` for summary rows only. |
| `sinceDate` | string | empty | Keep only reviews written on or after this date (`2026-01-31`). Reviews arrive newest first, so a listing stops as soon as it reaches older ones. |
| `minRating` | integer | empty | Keep only reviews of at least this many stars. |
| `maxRating` | integer | empty | Keep only reviews of at most this many stars — set it to `3` to read just the critical ones. |
| `languages` | list of strings | empty | Keep only reviews written in these languages (`en`, `fr`, `es`). Empty keeps every language. |
| `includeListingSummary` | boolean | `true` | Reads each listing's published rating breakdown, star histogram and topic tags and attaches them to that listing's summary row. Every listing gets a summary row either way; switched off, those rows carry only the measures derived from the reviews collected. |
| `includeTranslations` | boolean | `false` | Adds Airbnb's own translation of a review alongside the original text, where it publishes one. |
| `maxRequests` | integer | `1500` | Hard ceiling on how much work one run may do. The run finishes cleanly with whatever it has collected when the ceiling is reached. |
| Listings in parallel | integer | `4` | How many listings are worked on at the same time, up to 8. |
| `currency` | select | `USD` | Currency for any money shown alongside a listing. |
| `locale` | select | `en` | Language for listing titles, category labels and topic tags. |
| `proxyCountry` | select | Automatic | Country to appear to browse from. It does not change which reviews you get. |

### Output fields

#### Review rows (`recordType: "review"`)

| Field | Type | Description |
| --- | --- | --- |
| `reviewId` | string | Airbnb's own identifier for the review. |
| `listingId` / `listingUrl` | string | The listing the review belongs to, and a link to it. |
| `rating` | number | The guest's star rating, 1 to 5. |
| `reviewText` | string | Full review text, in the language the guest wrote it. |
| `reviewLanguage` | string | Language code of that text. Not the reviewer's nationality. |
| `createdAt` | date | When the review was written, to the second. |
| `localizedDate` | string | The month-and-year label Airbnb shows on the page. |
| `reviewerFirstName` | string | Reviewer's first name — the only name Airbnb publishes. |
| `reviewerId` / `reviewerContextualUserId` | string | The reviewer's two public identifiers. |
| `reviewerPictureUrl` / `reviewerProfilePath` | string | Reviewer's photo and public profile path. |
| `reviewerIsSuperhost` | boolean | Whether the reviewer is themselves a Superhost. |
| `reviewerTenureText` | string | How long the reviewer has been on Airbnb ("2 years on Airbnb"). This is tenure, **not** a location. |
| `reviewHighlight` / `highlightType` | string | Airbnb's short label for the stay, e.g. "Stayed a few nights". |
| `hostResponseText` / `hostResponseDateText` | string | The host's public reply and the month it was published, when there is one. |
| `hostId` / `hostFirstName` | string | The host who received the review. |
| `translatedText` / `translatedLanguage` / `needsTranslation` / `translationDisclaimer` | string / boolean | Airbnb's machine translation, when translations are switched on. |
| `reviewPhotoUrls` | array | Photos attached to the review, when the guest added any. |
| `scrapedAt` | date | When this run collected the review. |

#### Listing summary rows (`recordType: "listingSummary"`)

| Field | Type | Description |
| --- | --- | --- |
| `listingTitle` / `propertyType` / `personCapacity` | string / number | The listing's headline, its property type and how many guests it sleeps. |
| `bedrooms` / `beds` / `bathrooms` / `isStudio` | number / boolean | Layout as published on the listing's own summary line. |
| `overallRating` / `overallCount` | number | The headline rating and review count Airbnb shows. |
| `isGuestFavorite` / `isNewListing` | boolean | Airbnb's Guest Favourite badge, and whether the listing is too new to be rated. |
| `categoryRatings` | array | The six published category scores — cleanliness, accuracy, check-in, communication, location, value — as numbers. |
| `ratingHistogram` | array | The 1–5 star distribution, each band with its exact share. |
| `topicTags` | array | Airbnb's own review themes with the number of reviews mentioning each. |
| `reviewsRetrieved` / `sampleSize` | number | How many reviews this run read, and how many survived your filters. Every derived measure below is computed over `sampleSize`. |
| `sampleAverageRating` | number | Average star rating across the collected reviews. |
| `sampleFirstReviewAt` / `sampleLastReviewAt` / `sampleMonthsCovered` | date / number | The window the collected reviews span. |
| `publishedReviewCount` | number | The total Airbnb displays, recorded for comparison. |
| `hostResponseRate` | number | Share of collected reviews the host publicly replied to. |
| `medianResponseLagDays` | number | Typical gap between a review and the host's reply. Month-granular — see the limits below. |
| `responseMetricsFromSampleOnly` | boolean | Always true: reply measures describe the reviews collected, not the listing's whole history. |
| `reviewsPerMonth` | array | Review volume per calendar month, oldest first. The closest public signal for booking volume. |
| `monthlyRatingTrend` | array | Average rating per calendar month, oldest first. |
| `shareSub4` | number | Share of collected reviews rated below four stars. |
| `languageMix` | array | Languages the collected reviews were written in, commonest first, with shares. |
| `sectionsLookComplete` | boolean | False when Airbnb served the listing's page without its rating breakdown; the derived measures are still reported. |
| `reviewsStopReason` | string | Why collecting stopped for this listing — end of the reviews, your limit, your start date, or the run ceiling. |
| `scrapedAt` | date | When this run collected the listing. |

### Pricing

This actor is **pay per event**, with two charges and nothing else.

- **Review saved** — the built-in per-result charge, **$0.0003 per row**, which is
  **$0.30 per 1,000 rows**. It fires once for every row written to the dataset.
- **Listing rating summary** — **$0.006**, charged once for each listing whose
  published rating breakdown, star histogram and topic tags were actually
  retrieved. A listing that only produced reviews is never charged for it, and
  turning the summary off means it never fires at all.

A typical job — 50 listings with about 100 reviews each — is 5,000 review rows
plus 50 summaries, so **about $1.80** for the whole run. A single listing with 78
reviews costs about **$0.03**. Reviews that fail your date, rating or language
filters are never saved and therefore never charged, so a run that reads a
thousand reviews to keep the forty critical ones bills forty rows. If you set a
maximum cost for a run, the actor stops cleanly once that budget is spent instead
of overshooting it.

### Limits & what this actor cannot do

Most of these are Airbnb's own ceilings rather than this actor's, and all of them
are real:

- **Reviews arrive newest first, and that is the only order offered.** There is
  no "worst first" mode; use the rating filters to isolate critical reviews
  instead.
- **There is no ceiling on how many reviews can be collected for one listing.**
  Airbnb serves them in batches and the actor keeps going until it reaches the
  end, your per-listing limit, or your start date.
- **Host replies are published for only a small share of reviews.** On a typical
  listing a handful out of dozens carry one, so a low reply rate is usually the
  truth about the host rather than missing data.
- **The reviewer's location line is how long they have been on Airbnb, not where
  they live.** Airbnb does not publish a reviewer country anywhere in this data,
  and the actor does not invent one. The field is named `reviewerTenureText` for
  exactly that reason.
- **Review language is not the reviewer's nationality.** They are separate things
  and must not be conflated.
- **Reply timing is month-granular.** Airbnb publishes the reply date as a month
  and year only, so `medianResponseLagDays` is a comparative signal between
  listings, never a service-level measurement.
- **Every derived measure describes the reviews this run collected.** Reply rate,
  monthly volume, rating trend, sub-4-star share and language mix are computed
  over `sampleSize`, so they change if you narrow your filters or your
  per-listing limit. `overallRating`, `overallCount`, the category scores and the
  histogram are Airbnb's own published figures and cover the listing's whole
  history.
- **A narrow filter reads more than it saves.** With a start date, a rating band
  or a language list set, a listing keeps going until it has enough matching
  reviews — or until it reaches its equal share of the run ceiling, so that one
  listing can never use up the whole run and leave the others empty. Every
  summary row says why that listing stopped, and raising the ceiling lets a
  narrow filter go deeper.
- **Category ratings and the star histogram are the listing's published
  figures**, not something recomputed from the reviews — the two can differ
  slightly, and occasionally a listing's page is served without them, in which
  case `sectionsLookComplete` is false.
- **Bedroom, bed and bathroom counts come from the listing's own summary line**,
  so a listing that writes it unusually may leave one of them blank.
- **Finding listings from a market name is not exhaustive.** Airbnb shows only a
  few hundred listings for any single search and repeats itself past that point,
  so use a map area — or supply the room links yourself — when you need a
  specific set of properties.
- **Only English is verified end to end.** Other languages are best-effort: the
  reviews themselves are unaffected, but labels and titles may vary.
- **Speed depends on the size of the job and on how quickly Airbnb answers**, so
  no fixed throughput is promised.
- **Airbnb's terms prohibit automated access.** You are responsible for using
  what you collect lawfully and in line with the source site's terms, including
  any personal-data rules that apply where you operate.

### FAQ

**Do I need an Airbnb account?**
No. No account, no login, no personal details of any kind — paste a room link and
run it.

**Will it get every review a listing has?**
Yes, as long as your per-listing limit is high enough and you have not set a
start date or a rating filter. Compare `sampleSize` against
`publishedReviewCount` in the summary row to confirm it for any listing; the two
can differ by one or two because Airbnb's displayed total and its review list are
not always in perfect step.

**Can I run it on a schedule?**
Yes, and that is the cheapest way to track a listing over time. Set `sinceDate`
to the date of your last run: because reviews come back newest first, each
listing stops as soon as it reaches older reviews, so a daily or weekly run only
pays for what is new.

**Can I get only the bad reviews?**
Set `maxRating` to 3 (or 2). Filtered-out reviews are never saved, so you only
pay for the ones you keep.

**What if a listing has no reviews yet?**
You still get its summary row, with `sampleSize` of zero and null derived
measures rather than misleading zeros. Nothing is invented.

**Is the reviewer's country included?**
No. Airbnb does not publish it. The line that looks like a location is the
reviewer's time on Airbnb, and it is labelled as such.

**Can I collect reviews for a whole city?**
Partly. Give it a market name and it will find listings for you, but Airbnb only
shows a few hundred listings per search, so a whole-city sweep is not something
this actor can promise. For a defined comp set, supply the room links.

# Actor input Schema

## `listingUrls` (type: `array`):

The Airbnb listings to collect reviews for. Paste room links exactly as they appear in the address bar, for example "https://www.airbnb.com/rooms/5398166". Up to 500 listings per run.

## `listingIds` (type: `array`):

The same thing as the links above, but as bare numeric ids — handy when the ids come out of a spreadsheet or an earlier run.

## `query` (type: `string`):

A place to collect listings from when you have not supplied any links — type it exactly as you would in Airbnb's own search box, for example "Paris, France" or "Austin, Texas". Also required whenever you use a map area below.

## `bbox` (type: `object`):

Restrict the market above to a rectangle on the map, as {"neLat": 48.9, "neLng": 2.45, "swLat": 48.8, "swLng": 2.25} with the north-east corner first. Leave empty for the whole market. A map area must always be paired with a market name.

## `maxListings` (type: `integer`):

How many listings to work through in one run.

## `maxReviewsPerListing` (type: `integer`):

Upper bound on reviews collected for each listing. Reviews arrive newest first, so a smaller number gives you the most recent ones. Set it to 0 to collect only the per-listing summary row with the rating breakdown, histogram and topic tags.

## `sinceDate` (type: `string`):

Keep only reviews written on or after this date, as YYYY-MM-DD. Because reviews arrive newest first, the actor stops a listing as soon as it reaches older reviews — which makes scheduled runs cheap.

## `minRating` (type: `integer`):

Keep only reviews rated at least this many stars. Leave empty for all of them.

## `maxRating` (type: `integer`):

Keep only reviews rated at most this many stars — set it to 3 to read just the critical ones.

## `languages` (type: `array`):

Keep only reviews written in these languages, as short codes such as "en", "fr", "es". Leave empty to keep every language. A review's language is not the reviewer's nationality.

## `includeListingSummary` (type: `boolean`):

Reads each listing's own page for the published overall rating and review count, the six category ratings, the 1-5 star histogram, Airbnb's topic tags, the property type, guest capacity and bedroom/bed/bathroom counts, and attaches them to that listing's summary row. This is what the per-summary charge covers. Every listing gets a summary row either way — switched off, those rows carry only the figures derived from the reviews collected.

## `includeTranslations` (type: `boolean`):

Adds Airbnb's own translation of a review when it publishes one, alongside the original text. Roughly half of non-English reviews carry one.

## `maxRequests` (type: `integer`):

A hard ceiling on how much work one run may do. The projected cost is written to the log before any collecting starts, and the run finishes cleanly with whatever it has once the ceiling is reached.

## `maxConcurrency` (type: `integer`):

How many listings to work on at the same time. 4 is a good balance; lower it if you see failures.

## `currency` (type: `string`):

Currency used for any money Airbnb shows alongside a listing. Always set it explicitly so results stay comparable between runs.

## `locale` (type: `string`):

Language for listing titles, category labels and topic tags. Only English is verified end to end; other languages are best-effort.

## `proxyCountry` (type: `string`):

Which country the run should look like it is browsing from. It does not change which reviews you get; leave it on the default unless you have a reason to change it.

## Actor input object example

```json
{
  "listingUrls": [
    "https://www.airbnb.com/rooms/5398166"
  ],
  "listingIds": [],
  "query": "",
  "maxListings": 50,
  "maxReviewsPerListing": 200,
  "sinceDate": "",
  "languages": [],
  "includeListingSummary": true,
  "includeTranslations": false,
  "maxRequests": 1500,
  "maxConcurrency": 4,
  "currency": "USD",
  "locale": "en",
  "proxyCountry": ""
}
```

# Actor output Schema

## `dataset` (type: `string`):

No description

# API

You can run this Actor programmatically using our API. Below are code examples in JavaScript, Python, and CLI, as well as the OpenAPI specification and MCP server setup.

## JavaScript example

```javascript
import { ApifyClient } from 'apify-client';

// Initialize the ApifyClient with your Apify API token
// Replace the '<YOUR_API_TOKEN>' with your token
const client = new ApifyClient({
    token: '<YOUR_API_TOKEN>',
});

// Prepare Actor input
const input = {
    "listingUrls": [
        "https://www.airbnb.com/rooms/5398166"
    ],
    "maxListings": 50,
    "maxReviewsPerListing": 200,
    "maxRequests": 1500,
    "maxConcurrency": 4,
    "currency": "USD",
    "locale": "en"
};

// Run the Actor and wait for it to finish
const run = await client.actor("apt_marble/airbnb-reviews-scraper").call(input);

// Fetch and print Actor results from the run's dataset (if any)
console.log('Results from dataset');
console.log(`💾 Check your data here: https://console.apify.com/storage/datasets/${run.defaultDatasetId}`);
const { items } = await client.dataset(run.defaultDatasetId).listItems();
items.forEach((item) => {
    console.dir(item);
});

// 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/js/docs

```

## Python example

```python
from apify_client import ApifyClient

# Initialize the ApifyClient with your Apify API token
# Replace '<YOUR_API_TOKEN>' with your token.
client = ApifyClient("<YOUR_API_TOKEN>")

# Prepare the Actor input
run_input = {
    "listingUrls": ["https://www.airbnb.com/rooms/5398166"],
    "maxListings": 50,
    "maxReviewsPerListing": 200,
    "maxRequests": 1500,
    "maxConcurrency": 4,
    "currency": "USD",
    "locale": "en",
}

# Run the Actor and wait for it to finish
run = client.actor("apt_marble/airbnb-reviews-scraper").call(run_input=run_input)

# Fetch and print Actor results from the run's dataset (if there are any)
print("💾 Check your data here: https://console.apify.com/storage/datasets/" + run["defaultDatasetId"])
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item)

# 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/python/docs/quick-start

```

## CLI example

```bash
echo '{
  "listingUrls": [
    "https://www.airbnb.com/rooms/5398166"
  ],
  "maxListings": 50,
  "maxReviewsPerListing": 200,
  "maxRequests": 1500,
  "maxConcurrency": 4,
  "currency": "USD",
  "locale": "en"
}' |
apify call apt_marble/airbnb-reviews-scraper --silent --output-dataset

```

## MCP server setup

```json
{
    "mcpServers": {
        "apify": {
            "command": "npx",
            "args": [
                "mcp-remote",
                "https://mcp.apify.com/?tools=apt_marble/airbnb-reviews-scraper",
                "--header",
                "Authorization: Bearer <YOUR_API_TOKEN>"
            ]
        }
    }
}

```

## OpenAPI specification

Download the OpenAPI definition: https://api.apify.com/v2/acts/QuYcP87wWvHAC5imY/builds/vJhxy4ZkgilcXzq5v/openapi.json