Ximilar Visual Search
ximilar.com
- Free plan
- Web
- iPhone and iPad
- Android
- API
- Self-hosted
What Ximilar Visual Search is
Ximilar Visual Search uses AI-powered vector search to find similar images in a collection. It indexes visual descriptors for similarity searches across millions of images and supports metadata filtering. Images can be added through the API by URL or Base64 data, with up to 10 records per request. Search operations include visual search, multi-image queries, ranking by visual similarity, and object detection followed by similarity search. Listed uses include reverse image search, search by photo, similar product or graphic recommendations, and matching images across collections. The service is available through an API and web platform, and can be integrated with existing systems. Ximilar says it creates encoded image representations and discards the image files; collections may retain customer-supplied URLs. It lists a free plan with one collection of up to 1,000 images, and paid plans from $69/mo. Formats include JPG, JPEG, PNG, WEBP, HEIC, BMP, TIFF, JFIF, and GIF with first-frame-only support.
Who it suits
It may suit teams working with e-commerce, stock photos, industrial or scientific image collections, and security systems. API access and metadata filtering may be useful when integrating image search into an existing system.
Good to know, and what to watch
- Searches visual similarity across millions of images.
- Supports filtering by metadata.
- Accepts image URLs or Base64 data through API.
- Free plan includes one collection of up to 1,000 images.
- Custom plans offer on-premise and edge-device deployment.
- Free plan is limited to one collection.
- Free plan caps collections at 1,000 images.
- GIF support is limited to the first frame.
- Monthly API credits expire at the end of the billing period.
Plans
Every plan and priceProfessional 2M
Professional 3M
Professional 4M
Professional 5M
Free plan
Getting started
- Open the Ximilar Visual Search website
- Choose the Free plan or a paid plan
- Use the web platform or API
- Create a collection
- Add images by URL or Base64 data
- Authenticate API requests with the unique authorization token
Limits to check
The Free plan costs nothing, is billed forever, and includes 1,000 API credits per month, one collection, up to 1,000 images in that collection and three similarity tasks. Monthly credits expire at the end of their billing period; separately purchased credit packs do not expire.
The Club's review
Ximilar Visual Search is an API-driven service for finding visually similar images inside managed collections. It suits developers and organisations embedding image discovery in products, catalogues or internal systems. Its strongest advantage is the range of search operations; buyers wanting a simple consumer reverse-image tool should look elsewhere.
Overview
Ximilar converts images into encoded visual representations and searches those representations against a collection. The vendor says its indexing handles millions of images and supports metadata filtering, so results can combine visual resemblance with structured constraints. That makes the service relevant to e-commerce recommendations, stock-photo libraries, industrial and scientific collections, and security systems.
Images are added through the API by URL or Base64 data. A request can insert up to 10 records, which is manageable for incremental feeds but means bulk imports must be batched. Supported formats are JPG, JPEG, PNG, WEBP, HEIC, BMP, TIFF and JFIF; GIF processing uses only the first frame.
Key features
More than one search path
The API provides standard visual search, object detection followed by similarity search, multi-image queries and ranking by visual similarity. This gives developers several retrieval patterns instead of forcing every use case into a single-image lookup. Metadata filtering is particularly useful when visual matches still need to respect product, category or collection attributes.
Integration and data handling
API access lets teams connect Visual Search to an existing application rather than adopting a separate user interface. Ximilar says it analyses images into encoded representations and then discards the image files, although customer-supplied URLs may remain in collections. Custom training data is stored on Prague, EU servers; GDPR applies, and customer data is not used to train another customer's models. API communication uses a unique authorization token. Ximilar administrators can access images when asked, for example to resolve a problem, so organisations should account for that support access in their governance.
Pricing
The Free plan costs nothing, is billed forever and includes 1,000 API credits per month, one collection, up to 1,000 images in that collection and three similarity tasks. It is a useful proof-of-concept allowance, but the single collection and 1,000-image ceiling rule it out for a growing catalogue. Unused monthly credits expire at the end of each monthly period; separately purchased credit packs do not.
Business 100K costs $5,969 a month and provides 100,000 API credits, three collections and 100,000 images per collection. Business 300K costs $175,199 a month for 300,000 credits, five collections and 300,000 images per collection. Business 500K costs $285,329 a month for 500,000 credits, eight collections and 500,000 images per collection. These tiers fit organisations that need explicit, finite collection limits and higher monthly throughput.
Professional 1M costs $499,579 a month; Professional 2M costs $9,301,079 a month; Professional 3M costs $12,991,499 a month; Professional 4M costs $16,351,899 a month; Professional 5M costs $19,452,259 a month; and Professional 6M costs $22,292,589 a month. Each includes 10 collections, unlimited collection size and, respectively, 1,000,000 to 6,000,000 monthly API credits. Business includes standard support and one ML expert consultation, while Professional adds priority support and regular ML expert consultations. The pricing note says paid plans start at $69/mo, while the named high-volume plans carry the prices above; confirm the applicable quote before committing.
Platforms
Visual Search is offered through an API, web access and self-hosted deployment. Ximilar also lists on-premise and edge-device deployment through custom plans, giving regulated or latency-sensitive teams a route beyond the standard hosted tiers.
Who it's for
This is a strong candidate for e-commerce teams building similar-product recommendations, stock-photo operators, industrial and scientific image repositories, and security applications that need programmable matching. It also suits organisations with developers who can manage API ingestion and metadata. It is not the natural choice for an individual who wants an off-the-shelf reverse-image lookup with no integration work.
Pros and cons
- Pros: similarity search scales to collections described as millions of images, with metadata filtering for more precise retrieval.
- Pros: object-detection searches, multi-image queries and visual ranking support more advanced workflows than a basic reverse-image search.
- Pros: API, self-hosted and custom on-premise or edge options accommodate different deployment requirements.
- Pros: EU storage, GDPR coverage and a stated separation of customer data from other customers' model training support privacy-conscious deployments.
- Cons: the Free plan is limited to one 1,000-image collection, 1,000 monthly credits and three similarity tasks.
- Cons: monthly credits expire, and API insertion batches are capped at 10 records, adding operational overhead for large imports.
- Cons: administrator access to images can occur when support is requested, which may require internal approval procedures.
Alternatives
For a broader shortlist, see AI Reverse Image Search Tools. Choose Berify instead when a web-based reverse-image service and small, clearly priced image tiers are sufficient. infrimage is an option for web searches organised into projects, with a 10-day free period. Google Lens is the simpler free choice across mobile, desktop and web platforms. ImageRights Discovery suits a free pilot covering the first 100 images. Pick Imagga Visual Search when you also need its listed background-removal, barcode-recognition and OCR APIs. trace.moe is a free, donation-funded option focused on anime image searching. Watson offers a small, credit-based mobile search service. PixGuard is aimed at image scanning through web, extension, API and self-hosted access.
Verdict
Choose Ximilar Visual Search if you need programmable similarity search across a substantial image collection, especially with metadata filters, multi-image queries or object-detection workflows. The free tier is a sensible evaluation route, but its collection and credit caps are restrictive, and larger plans require a serious budget. Look elsewhere if your priority is a ready-made reverse-image search for occasional personal use.
The verdict
Ximilar Visual Search combines image similarity search with API operations such as multi-image queries and object detection. The free plan has clear collection limits, while deployment options include on-premise and edge devices through custom plans.
Questions people ask
Is there a free plan?
Yes. The Free plan is free, billed forever, with one collection and 1,000 API credits per month.
How much do paid plans cost?
Paid plans are listed from $69/mo. The pricing page also lists Business and Professional plans with different credit and collection allowances.
Which platforms are supported?
Ximilar lists API, self-hosted and web access. On-premise and edge-device deployment are available through custom plans.
What image formats are supported?
Supported formats are JPG, JPEG, PNG, WEBP, HEIC, BMP, TIFF and JFIF. GIF is supported using only its first frame.
Can it integrate with existing systems?
Yes. Ximilar provides API access for integration with existing systems, including image uploads by URL or Base64 data.
How are images handled?
For visual search, images are converted into encoded representations and then discarded. Customer-supplied URLs may remain in collections.
What the maker publishes
- Access controls
- API communication uses a unique authorization token, and Ximilar administrators may access images after being asked, such as when resolving a problem. source
- API endpoints
- Visual-search endpoints cover photo, image matching, product, custom, fashion and home-decor similarity services. source
- API operations
- The API includes insert, delete, get, find, update, visual KNN, object-detection KNN, multi-image search, ranking, range search and near-duplicate operations. source
- Authentication
- Requests use an API token in the Authorization header and a collection identifier in the collection-id header. source
- Batch limit
- The API supports inserting up to 10 records per request. source
- Batch limit
- Insert requests can contain at most 10 records, with each image billed separately according to its operation. source
- Capterra listing
- Capterra lists Ximilar with a 0.0 rating and no published user reviews. source
- Company size
- LinkedIn lists Ximilar as a privately held company with 11–50 employees. source
- Compliance
- Ximilar states that it complies with GDPR and lists ISO 27001, ISO 9001, ISO 14001, ISO 45001 and ISO 20000 certifications. source
- Credit billing
- API calls consume credits according to method complexity, while free plans receive monthly credits and paid plans offer configurable supplies. source
- Credit expiry
- Unused monthly-plan credits expire at the end of the monthly period, while purchased credit packs do not expire. source
- Data handling
- For visual search, Ximilar says it analyzes images to create encoded representations and then discards the image files; collections may retain URLs supplied by the customer. source
- Deployment
- Ximilar lists on-premise and edge-device deployment as options available through custom plans. source
- Flows
- Flows chain classification, tagging, regression, detection and similarity models into one REST endpoint and support conditional branches and asynchronous processing. source
- Format limitation
- AVIF, JP2, PSD, SVG, HEIF and PDF are listed as unsupported input formats. source
- Founded
- LinkedIn lists Ximilar as founded in 2016. source
- Headquarters
- Brno, Czech Republic source
- Image retention
- Ximilar says inference and visual-search images are not stored; only extracted metadata is retained, while training uploads are stored in Amazon S3. source
- Input methods
- Images can be added to a collection through the API using a URL or Base64 data. source
- Integrations
- Ximilar says its Visual Search platform is accessible through an API and can be integrated with existing systems. source
- Legal name
- Crunchbase identifies the company as Ximilar s.r.o., headquartered in Brno, Czech Republic. source
- Limit: deprecated endpoints
- The older fashion and home-decor similarity endpoint paths are deprecated and scheduled for removal. source
- Limit: public image URLs
- When sending images by URL, the link must be a publicly accessible direct file URL rather than an HTML page or private preview. source
- Limit: sparse independent feedback
- G2 says there are not enough reviews to provide buying insight, and GetApp lists no user reviews or user-derived ratings. source
Checked against the maker's pages on 15 Oct 2026.