Index all profiles and casts on the Farcaster protocol using Warpcast APIs. Powers Searchcaster, Fardrop and others.

Soon, both profiles and casts will read from Farcaster Hubs instead of client APIs.

Requirements: Node.js, Yarn, Docker, Supabase CLI

In the project directory, create a local Supabase instance. This will create all the tables for you.

supabase start

Rename .env.example to .env and configure your variables with the credentials generated from the previous step. Your SUPABASE_URL will be the API URL from the terminal output. The Studio URL is not necessary, but you may want to use it to view your database tables.

cp .env.example .env

If you don't have a Merkle auth token yet, set the FC_MNEMONIC environment variable to your Farcaster recovery phrase and run the following command to generate a token.

yarn install
yarn run auth

Seed your database with protocol data. This will take ~5-10 minutes for profiles and casts (default), or ~30 minutes for everything (include the --verifications flag).

yarn run seed
# or
yarn run seed --verifications

The Merkle APIs don't include a registration timestamp for users. For new registrations, we get the timestamp by watching events on the ID Registry contract. If you were running the previous version of this indexer, you can migrate this data. Otherwise you can skip this step.

yarn run migrate

Run the indexer

yarn start

Postgres full text search is a lot more performant and robust than pattern matching, especially when querying the casts table. It's a powerful search engine that can:

• stem words (e.g. "run" matches "runs", "running", and "ran")
• ignore stop words (e.g. "the" and "a")
• weight and rank results

The data can be queried with SQL or the Supabase client. For example, the following code will match casts that contain either "farcaster" and "warpast" OR "activitypub" and "mastodon".

SELECT *
FROM casts
WHERE fts @@ to_tsquery('english', '(farcaster & warpcast) | (activitypub & mastodon)')

supabase
  .from('casts')
  .select()
  .textSearch('fts', '(farcaster & warpcast) | (activitypub & mastodon)')

See full text search on Supabase for more details.

Create an empty Supabase project and connect to the CLI. If you get a warning that says "Local config differs from linked project", update the major_version in supabase/config.toml to 15.

supabase login
supabase link --project-ref <project-id>

Push your database schema

supabase db push

I recommend hosting the indexer on Railway.