Meet Anton.
Your search evaluation copilot.
Anton is a specialized AI agent for programmatically evaluating search results. Anton takes a search query and a JSON object, and returns a “great”, “ok”, or “bad” judgement along with an explanation for it’s decision.
Try it live
“The search result appears to show a coat that is yellow in color ('colour_group_name') and has a solid pattern ('graphical_appearance_name'), which aligns with the color aspect of the user's query for a 'yellow rain coat.' The image provided also directly...”
Anton speaks normal, RESTful HTTP and works asynchronously. So whether you’re doing some quick local development or batching thousands of judgements for daily monitoring, you can pass the workload to Anton and retrieve responses when they’re available.
A platform that reads and sees, together.
True multimodal understanding that interprets semantic meaning in text and crawls images in your Object to build a holistic understanding.
Most datasets in real production search environments involve hybrid data — strings, numbers, and URLs. Objective knows how to interpret each.
Anton is built on Objective’s foundational intent understanding, taking advantage of all of the semantic processing and multimodal reasoning that come out of the box with everything on the Objective Platform. Each evaluation also returns human-readable explanation that can be programmatically parsed and summarized for aggregate analysis.
Evaluate any search installation. Or compare them all.
Whether you currently have a highly customized with Algolia™ or ElasticSearch™, or you’ve built your own home-grown search system, evaluating the relevance of that system is as easy as passing results for queries you’re interested in to Anton for judgement.
Or compare two completely different search systems - load your data into both, and decide for yourself which platform returns more relevant results for your dataset and your users.
Build a whole new breed of search observability.
You can’t improve what you can’t measure. Your dataset changes constantly, which means your search relevancy does too. What if you could monitor search relevancy in realtime, as your system changes?
Build a daily worker process that takes search results for your top queries, evaluate them with Anton, and roll them into your internal reporting surfaces.
Pricing that scales with you.
Anton is priced to scale with your needs, including a free tier of usage to get you started experimenting, and per-request pricing that scales with you.
Built on a modern, AI-native search platform.
Anton runs on the same AI-native platform that all of Objective is built on. With semantic understanding and multimodal search as core building blocks, Anton is ready to handle any kind of evaluation task you can throw at it.
