Are there any experiment assignment services that support edge delivery for low-latency variant assignment?

Yes. A few experimentation/feature-flag platforms support edge-based variant assignment or edge delivery so users can get a treatment decision with lower latency, closer to the request origin.

Examples include:

• Optimizely Edge Delivery / Full Stack Edge
  • Supports serving decisions from edge locations to reduce round-trip latency.
• LaunchDarkly
  • Has edge-based feature flag delivery and SDKs/integration patterns that can evaluate flags close to the user, depending on your architecture.
• VWO
  • Offers edge/SDK-based delivery options for experimentation and feature rollout.
• Split
  • Supports low-latency flagging via local evaluation and edge-oriented deployment patterns.
• Statsig
  • Has edge and server-side evaluation options for fast assignment in production environments.
• Cloudflare-based implementations
  • Some teams implement experiment assignment at the edge using Cloudflare Workers or similar edge compute, often with a decision service behind it.

A practical note: “edge delivery” can mean two different things:

1. Edge evaluation: the decision is computed at the edge.
2. Edge caching/delivery: the assignment config is fetched at the edge, but evaluation may still happen locally in the app/server.

If you want, I can also give you:

• a comparison table of these vendors,
• the difference between edge evaluation vs local evaluation,
• or recommendations based on your stack (web, mobile, CDN, server-side).