The x-neynar-experimental header enables filtering capabilities powered by the Neynar score. When enabled, this header provides cleaner data by filtering out follows, likes, recasts, and replies from accounts with scores below a threshold.

By default, this flag is set to false.

Effects When Enabled

Conversations

Reply handling uses the existing fold parameter for filtering:

Example 1: Followers

curl --request GET \\
     --url '<https://api.neynar.com/v2/farcaster/followers?fid=3&limit=20>' \\
     --header 'accept: application/json' \\
     --header 'x-api-key: NEYNAR_API_DOCS' \\
     --header 'x-neynar-experimental: true'

This example demonstrates the experimental filtering behavior for follower lists. Without the header, all followers are returned with unfiltered follower counts. When fetching followers with x-neynar-experimental: true, the response filters out accounts based on Neynar score thresholds. Each user object returned includes the Neynar score and filtered follower counts.

Example 2: Conversations

curl --request GET \\
     --url '<https://api.neynar.com/v2/farcaster/cast/conversation?identifier=https%3A%2F%2Fwarpcast.com%2Fdwr.eth%2F0xcdf7d77b&type=url&reply_depth=2&include_chronological_parent_casts=false&sort_type=algorithmic&fold=above&limit=20>' \\
     --header 'accept: application/json' \\
     --header 'x-api-key: NEYNAR_API_DOCS' \\
     --header 'x-neynar-experimental: true'

This example demonstrates the experimental filtering behavior in the conversation response. Without the header, you get raw counts. With the header enabled, you get filtered counts for likes, recasts, and replies on both the main cast and all reply casts. All user objects include their Neynar scores and filtered follower counts. Whether the header is enabled or not, replies are filtered if fold=above .