Onton

Onton is a next-generation search and discovery engine. We’re a tight-knit San Francisco-based team building the future of shopping.

We've built Ontology 1, a neurosymbolic model that answers complex, conversational, multimodal product queries more accurately than Amazon or Google Shopping, despite having indexed only 1% of their catalog.

Best in-class precision with smaller datasets

In the benchmark below, Ontology 1 wins 52 of 90 searches outright. (Google wins 19, and Amazon 16.) Ontology achieves 63.0% accuracy in its top 10 results, compared to Google's 54.3% and Amazon's 46.9%. We also compare multimodal queries. Amazon lacks multimodal search support, and Google, while it returns results, routinely misses the query. As far as we're aware, Onton is the only platform that handles complex, intent-heavy multimodal queries.

Methodology

We introduce Subtext-Decor-90, in which three independent multimodal judges — Claude Opus 4.8, Gemini 3.1 Pro, and GPT-5.52 — evaluate 90 text queries against Onton, Amazon, and Google Shopping, scoring the top 10 results from each engine. (Complete code and data may be found here.)

Subtext-Decor-90 was curated to include intent-heavy queries with rich aesthetic descriptors, negation, cultural references, and emotional framing. These are cases that resist traditional keyword and vector retrieval. They also reflect where product search is heading. In 2025, Bloomreach found that 57% of shoppers have used AI to help them shop, and 41% now search using natural language rather than keywords. A 2026 study by Klaviyo similarly found that consumers are abandoning keywords in favor of full phrases and questions.

For each query in Subtext-Decor-90, the judges saw all three result screenshots in a single call. Each judge scored the first 10 visible ranked result cards in each screenshot (left-to-right, top-to-bottom). P@10 was aggregated as the mean across judges, with 10,000-resample bootstrap CIs over queries. We report precision rather than recall, which is unanswerable here without full access to Amazon and Google’s catalogs.

Image and multimodal searches are excluded from Subtext-Decor-90 because a 1:1 comparison with Amazon and Google isn't possible. Amazon Lens is built around finding live products with a mobile camera, and doesn’t support multimodal queries at all. Google Lens supports multimodal, but unlike Google Shopping, doesn’t exclusively return products. We provide a separate Onton-vs.-Google comparison on 10 image and multimodal queries below.

Analysis

The judges grade on different scales: Gemini 3.1 Pro is harshest (mean P@10 0.43 across engines), Opus 4.8 the most generous (0.70), and GPT-5.5 sits in the middle (~0.51).

While Krippendorff's alpha across the three judges is 0.465, all three judges put the engines in the same order: Onton, then Google, then Amazon. This finding is robust across judges: Onton outperforms the next-best engine by a clear margin under every judge.

Failure Cases

Ontology’s failure cases are concentrated in functional spec-related queries where Amazon’s category metadata dominates. A few examples of Ontology 1 failure cases:

• “lamp that won't wake my partner if I read at 3am”
• “something to put on a weirdly deep windowsill”

Performance by search engine

Amazon does well on functional-spec queries where there's a clean keyword-to-category match (dimensions, materials, fixed attributes), and falls apart on aesthetic modifiers and stacked negation. Google Shopping does well when it can pull up a curated section for a structured constraint, and misses on subtle intent and explicit negation. Onton does well on the queries where you have to model what the person actually means instead of matching keywords.

Image and multimodal search

Below we compare 10 image and multimodal queries on Onton and Google. Google surfaces visually relevant results that often miss the query.

Looking forward

Next, we want to harden the benchmark and broaden its scope.

Hardening

• Richer ranking metrics: Adding NDCG@10 on graded (rather than binary) relevance judgments, would reward both correct ordering and degree of quality.
• Judge normalization: Normalizing scores before aggregating would tighten confidence intervals.
• Live, longitudinal evaluation: A one-shot screenshot benchmark is a snapshot, and search engines change. We can re-run it on a regular cadence.

Broadening

• Larger, broader query sets: Scaling up the number of queries would increase benchmark confidence.
• More engines: Extending to platforms like Pinterest, Wayfair, and Etsy would give a broader view of the search landscape.
• Coverage estimates: Providing a first, bounded read on coverage alongside our precision numbers.

Catalog and vertical expansion: Growing catalog breadth — and moving beyond home decor — is the clearest path to closing the functional-spec gap with Amazon and Google.

Accelerating the learning loop: Tightening that loop is the lever we expect to move our numbers most between benchmark runs.