Fuzzy Search Dataset Actor

Search any Apify dataset using typo-tolerant fuzzy matching. Point this Actor at an existing dataset, provide a query, and get back ranked results — even when spellings are imperfect. Try it directly in Apify Console.

What does Fuzzy Search Dataset do?

This Actor loads records from any Apify dataset and runs a fuzzy full-text search across one or more fields. It handles typos, partial matches, and word-order variations automatically. For example, searching "iphon pro mx" can still return results for "iPhone 15 Pro Max".

It's ideal for post-processing scraped data — after collecting a large dataset with another Actor, use this one to instantly build a search layer on top of it without any external infrastructure.

Why use Fuzzy Search Dataset?

• No search engine needed — works directly on any Apify dataset without Elasticsearch, Algolia, or similar tools
• Typo-tolerant — handles misspellings, abbreviations, and partial queries out of the box
• Multi-field search — search across title, description, brand, nested fields like product.name, or any combination
• Tunable relevance — control strictness, field weights, and minimum match length to fit your data
• Automation-ready — trigger via API, schedule it, or chain it after a scraping Actor in a workflow

How to use Fuzzy Search Dataset

1. Run a scraping Actor to build the source dataset (or use any existing dataset you already have in Apify Console).
2. Copy the dataset ID from the dataset URL or the Storage section of Apify Console.
3. Open this Actor and paste the dataset ID into the Dataset ID field.
4. Enter your search query and choose which fields to search.
5. Run the Actor — results are written to its output dataset, ranked by relevance score.

Input

Configure the Actor in the Input tab or pass a JSON object via the API.

Field	Type	Required	Default	Description
datasetId	string	✅	—	ID of the Apify dataset to search
query	string	✅	—	Text to search for
fields	array of strings		["title"]	Dataset fields to search (supports dot notation for nested fields)
limit	integer		20	Maximum number of results to return (1–1000)
threshold	number		0.35	Fuzzy strictness — 0.0 = exact only, 1.0 = match anything
ignoreLocation	boolean		true	Allow matches anywhere in the text, not just at the start
minMatchCharLength	integer		2	Minimum characters in a token before it counts as a match
includeScore	boolean		true	Attach a relevance score to each result (lower = better match)
includeMatches	boolean		false	Include matched text ranges — useful for keyword highlighting
extendedSearch	boolean		false	Enable advanced query syntax (^starts-with, !exclude, =exact)
weights	object		—	Per-field importance weights, e.g. {"title": 0.7, "description": 0.3}

Example input:

{
    "datasetId": "UoYaa1QjGdgdJrSHA",
    "query": "iphon pro max",
    "fields": ["title", "description"],
    "limit": 10,
    "threshold": 0.35,
    "weights": {
        "title": 0.8,
        "description": 0.2
    }
}

Output

Results are pushed to the Actor's default dataset as a single object containing the query, total result count, and an array of ranked matches.

Example output:

{
    "query": "iphon pro max",
    "totalResults": 3,
    "results": [
        {
            "rank": 1,
            "score": 0.04,
            "item": {
                "title": "Apple iPhone 15 Pro Max",
                "description": "6.7-inch Super Retina XDR display, A17 Pro chip",
                "brand": "Apple",
                "price": 1199
            }
        },
        {
            "rank": 2,
            "score": 0.18,
            "item": {
                "title": "iPhone 14 Pro Max",
                "description": "48MP main camera, Dynamic Island",
                "brand": "Apple",
                "price": 999
            }
        }
    ]
}

You can download results in JSON, CSV, Excel, or HTML from the dataset tab in Apify Console or via the Apify API.

Output data fields

Field	Format	Description
query	text	The search query that was used
totalResults	number	Number of results returned
results[].rank	number	1-based position in results (1 = best match)
results[].score	number	Relevance score (0.0 = perfect match, 1.0 = no match)
results[].item	object	Full original record from the source dataset

Tips and advanced options

Tuning the threshold:

• 0.2 — strict, good for product codes and exact names
• 0.35 — balanced (default), works well for product titles and descriptions
• 0.6 — loose, useful for free-text fields or short queries

Multi-field search with weights:

To boost title matches above description matches, set weights to {"title": 0.8, "description": 0.2}. Weights must sum to 1.0 across all searched fields.

Advanced query syntax (when extendedSearch is enabled):

Syntax	Meaning	Example
=iphone	Exact match	=iPhone 15
^apple	Starts with	^Apple
!samsung	Exclude	!Samsung
'pro	Includes token	'pro

Performance: The Actor loads the entire dataset into memory. For datasets over 100k records, consider filtering the source dataset first or increasing the Actor's memory allocation.

Pricing / Cost estimation

This Actor processes data in-memory without using a browser or proxy, so compute costs are low. A typical run over 10,000 records completes in under 30 seconds. Apify provides a free tier sufficient for many use cases.

FAQ and support

Is this legal? This Actor only reads from datasets you own or have access to on the Apify platform. It does not scrape any external websites.

Can I use this with datasets from other Actors? Yes — as long as you have access to the dataset ID, this Actor can read it.

The results are empty or not what I expected. Try raising the threshold value (e.g. to 0.5), enabling ignoreLocation, or adding more fields to the fields array.

Found a bug or want a feature? Open an issue in the Issues tab on this Actor's page. Custom solutions are also available — reach out via the Apify platform.

Resources

• Apify SDK documentation
• Apify Platform documentation
• Join the Apify developer community on Discord