Civitai Models Scraper

Civitai Models Scraper: AI image models search, downloads, and ratings

This Civitai models scraper helps you quickly explore AI image models, download counts, and ratings. Every run creates a clean dataset ready to export in JSON, CSV, or Excel formats.

Search for models, look up specific items by ID, browse recent image feeds, or grab full tag collections. Exported rows capture essential details like creators, commercial usage rights, latest version names, base models, and model page links.

Use cases

• Model research: Look up checkpoints, LoRAs, and styles while keeping download numbers right next to user ratings.
• Version tracking: Pass model IDs to keep track of new version releases, base architecture, and update trends.
• Image inspiration: Pull up fresh community creations or look up sample images linked to a specific model.
• Tag analysis: Extract tag lists to see how creators categorize their work across different styles.
• Scheduled runs: Set up recurring runs to automatically collect updated statistics over time.

What data does this actor extract?

Model listings record model IDs, titles, categories, content flags, creator usernames, tags, download totals, thumbs-up ratings, commercial permissions, latest version names, base model versions, and page links. Image feeds collect image URLs, dimensions, and timestamps. Tag runs output tag labels and link paths.

Field naming includes convenient camelCase and snake_case aliases (like modelType and model_type), making the data easy to work with in spreadsheets or custom data pipelines.

Please give credit to Civitai when sharing or republishing extracted datasets.

Input

Parameter	Type	Default	Description
mode	string	search_models	Search models, look up model IDs, pull image feeds, or list tags.
query	string	anime	Search term for model searches and tag filtering.
modelIds	array	4201	List of model IDs for direct lookup or image filtering.
types	array	empty	Model types to filter search results, such as Checkpoint or LORA.
sort	string	Highest Rated	Order by Highest Rated, Most Downloaded, or Newest.
period	string	AllTime	Time window: AllTime, Year, Month, Week, or Day.
includeNsfw	boolean	false	Enable to include adult/NSFW flagged content in results.
maxItems	integer	100	Maximum rows to collect in one run (up to 1000).
requestDelaySecs	number	1	Pause between calls in seconds.
userAgent	string	BigiByte research UA	Custom browser header identifier.
timeoutSecs	integer	300	Maximum run execution time in seconds.
requestTimeoutSecs	integer	30	Timeout per network request.
proxyConfiguration	object	No proxy	Custom proxy settings if needed.

Example input

{
    "mode": "search_models",
    "query": "anime",
    "modelIds": ["4201"],
    "types": ["LORA"],
    "sort": "Highest Rated",
    "period": "AllTime",
    "includeNsfw": false,
    "maxItems": 100,
    "requestDelaySecs": 1,
    "userAgent": "B2BLeadGenerationResearch/1.0 (contact: research@b2bleadgeneration.com)",
    "timeoutSecs": 300,
    "requestTimeoutSecs": 30,
    "proxyConfiguration": { "useApifyProxy": false }
}

Output

{
    "id": 16014,
    "name": "Anime Lineart / Manga-like (线稿) Style",
    "modelType": "LORA",
    "model_type": "LORA",
    "nsfw": false,
    "creator_username": "CyberAIchemist",
    "creatorUsername": "CyberAIchemist",
    "tags": ["lineart", "art style"],
    "downloadCount": 207410,
    "thumbsUpCount": 21429,
    "allowCommercialUse": [],
    "commercialUse": [],
    "latestVersion_name": "v3.0 manga-like",
    "latestVersionName": "v3.0 manga-like",
    "latestVersion_baseModel": "SD 1.5",
    "baseModel": "SD 1.5",
    "downloadUrl": "[https://civitai.com/api/download/models/28907](https://civitai.com/api/download/models/28907)",
    "url": "[https://civitai.com/models/16014](https://civitai.com/models/16014)",
    "scraped_at": "2026-10-12T00:00:00+00:00",
    "scrapedAt": "2026-10-12T00:00:00+00:00"
}

Field	Type	Description
id	integer	Unique model or image identifier.
name	string	Model title or tag label.
modelType	string	Model category (e.g., Checkpoint, LORA).
nsfw	boolean	Content safety flag.
creator_username	string	Profile name of the creator.
tags	array	Tags assigned to the model.
downloadCount	integer	Total downloads recorded on the listing.
thumbsUpCount	integer	Positive ratings count.
allowCommercialUse	array	Permissions for commercial use.
latestVersion_name	string	Name of the most recent version update.
latestVersion_baseModel	string	Base architecture (such as SD 1.5 or SDXL).
downloadUrl	string	Direct link to the model file download page.
url	string	Link to the main listing on Civitai.
scraped_at	string	Timestamp when the item was saved.
imageUrl	string	Direct link to the image on sample rows.

How it works

1. Reads your configuration parameters and prepares the extraction tasks.
2. Formats requests based on chosen search terms, model categories, and filters.
3. Automatically throttles requests using your specified delay (default is 1 second) to maintain smooth operations.
4. Handles temporary network glitches or busy server responses with retry delays.
5. Pages through results until your dataset reaches maxItems or ends cleanly.
6. Eliminates duplicate listings automatically and saves output in batches.

Limits and guidelines

Request speeds default to one call per second to keep runs smooth and reliable.

The scraper focuses strictly on metadata, ratings, and image links. It does not download actual model weight files. Always ensure you follow Civitai's content terms and credit creators when sharing findings.

Pricing

This actor does not include any hidden software surcharges. Standard Apify compute usage and data storage costs apply. maxItems defaults to 100 rows per run and supports up to 1000 rows.

FAQ

Does it download the actual AI model files?

No, weight files can be several gigabytes each. The actor extracts the metadata and direct download URLs so you can review options before downloading manually.

What happens if a model ID isn't found?

If a specific model ID returns an error or no longer exists, the scraper logs the issue into an error row and keeps processing the rest of your list.

How many rows can I export per run?

By default, runs collect up to 100 items. You can set maxItems higher depending on your needs, up to a maximum limit of 1000 items per run.

---

Trademark Notice

Civitai and associated brand names, logos, and trademarks belong to their respective owners. They are referenced here purely for identification purposes and do not imply any sponsorship or partnership.

Disclaimer

This tool is an independent scraper built to help users organize publicly accessible data. Users remain responsible for complying with site terms, rate limits, and licensing restrictions when collecting or republishing content.

Is web scraping legal?

This actor only collects publicly available data without circumventing paywalls or restricted areas. Scraping public content is widely practiced, but how you handle the data matters:

• Personal data: When collecting profile usernames or creator details, comply with relevant data protection guidelines such as GDPR or CCPA.
• Copyright: Respect creator rights. Avoid rehosting copyrighted model images or text descriptions without proper permission or attribution.
• Terms of use: Play fair by keeping request delays reasonable and honoring creator usage licenses.

To learn more, read Apify's guide: Is web scraping legal?