Palette is a front-end performance observability platform that ties slow user experiences—such as page loads, typing, and interactions—directly to the specific lines of code causing them. It collects end-user traces, JavaScript profiles, and metrics via a low-overhead SDK, enabling engineering teams to identify and fix regressions in minutes rather than months. The tool has been adopted by companies like Notion, which used it to reduce page load latency by 15-20% and typing latency by 15%.
Funding
€3M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
DKJBJGJG+7Founders
Product
Problem
Engineering teams often struggle to identify the root cause of front-end performance regressions because local profiling tools cannot reproduce the variability of real user hardware, flows, and document sizes. Reproducing and attributing slow experiences to specific code changes can take weeks or months, leading to unresolved issues and baselining metrics to a regressed state.
Solution
Palette is a performance observability platform that collects interaction latency, page load data, and JavaScript code profiles directly from end-user sessions, tying slow experiences back to the exact lines of code responsible. Its SDK runs at the lowest priority, minimizes main-thread blocking, and uses the Profiler WebAPI for low-overhead sampling, keeping total overhead to at most 3%. The platform captures traces, metrics, and profiles during key events like typing, scrolling, and clicks, then presents the data in a dashboard that helps engineers pinpoint regressions and optimization opportunities. This approach enables teams to move from identifying an issue to attributing it to a specific commit in minutes, significantly reducing time-to-resolution.
Target Audience
Primary customers are front-end engineering teams and engineering managers at web-based product companies who need code-level visibility into user-facing performance issues and regressions.
Features
- End-user session profiling that functions like a Chrome DevTools profiler powered by real user sessions, not synthetic tests
- Low-overhead SDK using the Profiler WebAPI with configurable sampling intervals, keeping total performance impact under 3%
- Collection of interaction metrics, web vitals, custom metrics, React performance metrics, and markers (tracing) for comprehensive visibility
- Privacy-focused data collection that captures only performance-relevant information, such as keypress latency without the key value
- Network payloads compressed asynchronously off the main thread and sent only during idle periods to minimize impact
- Support for browser, mobile, and desktop applications with configurable sampling rates for stable metric collection