Farcaster Casts + Profiles Scraper

Extract casts (posts) and profiles from Farcaster — the fast-growing decentralized social network — using keyless public APIs. No account, no API key, no credentials, no proxy spend. Public, logged-out data only.

Farcaster is the leading web3-native social protocol (the network behind the Warpcast / Farcaster app, backed by a16z). It is where crypto founders, builders, and communities publish in the open — a high-signal, under-scraped data source for AI pipelines, market research, and community intelligence.

What it does

• Search casts by keyword or phrase — full-text search across all public casts, newest first, with resolved authors and engagement counts
• Scrape profiles by username or FID — rich account data: bio, follower / following counts, verified addresses, connected X (Twitter) accounts
• Include user timelines — optionally pull each profile's recent casts in the same run
• Channel feeds — collect recent casts from any channel (e.g. /base, /degen, /farcaster) via the canonical Farcaster hub
• Unified, flattened output — casts and profiles land as clean, one-object-per-row dataset items with consistent field names

Why use this actor

Alternative	Problem	This actor
Neynar / other hosted Farcaster APIs	Require a paid API key + signup	Keyless — zero credentials
Generic "web3 social" scrapers	Don't cover Farcaster casts/profiles	Purpose-built for Farcaster
Raw Hubble hub calls	Return FID-only protobuf; no text search; no resolved authors	Full-text search + resolved usernames, display names, engagement
DIY scripts	Break on rate-limits / upstream blocks	Born hardened — never crashes, degrades gracefully

At publish time there is no dedicated Farcaster casts+profiles scraper on the Apify Store — this fills a clear gap for a network with millions of casts.

Modes

1. Search casts

{ "mode": "search", "searchTerms": ["ethereum", "base app"], "maxResults": 500 }

Full-text search over public casts. Each term is a separate query. Cursor pagination walks real depth.

2. Profiles

{ "mode": "profile", "usernames": ["dwr", "vitalik.eth"], "includeUserCasts": true, "maxResults": 100 }

Rich profiles by username (or fids: [3, 5650]). With includeUserCasts, also pulls each user's recent cast timeline.

3. Channel feed

{ "mode": "channel", "channels": ["base", "degen", "farcaster"], "maxResults": 300 }

Recent casts in a channel, pulled from the canonical Farcaster hub and joined to resolved author usernames.

Output fields

Casts

Field	Type	Description
hash	string	Canonical cast hash (unique id)
url	string	Permalink
text	string	Cast body text
timestamp	string	ISO 8601 timestamp
author_fid	integer	Author's Farcaster ID
author_username	string	Author @username
author_display_name	string	Author display name
author_follower_count	integer	Author followers
reply_count	integer	Replies (search/profile modes)
reaction_count	integer	Likes (search/profile modes)
recast_count	integer	Recasts (search/profile modes)
channel_key	string	Channel the cast was posted to
image_urls	array	Embedded image URLs
embed_urls	array	Embedded link URLs
has_media	boolean	Whether the cast has media
mentions	array	Mentioned users
parent_hash	string	Parent cast hash (if a reply)

Profiles

Field	Type	Description
fid	integer	Farcaster ID
username	string	@username
display_name	string	Display name
url	string	Profile link
bio	string	Bio text
follower_count	integer	Followers
following_count	integer	Following
pfp_url	string	Profile picture URL
external_url	string	Linked website
connected_accounts	array	Linked accounts (e.g. X/Twitter)

Common use cases

• Crypto & web3 research — track narratives, tokens, and protocols as builders discuss them in the open
• Brand & founder monitoring — follow mentions of your product, company, or people on Farcaster
• Community intelligence — monitor channels like /base or /degen for emerging topics
• AI / RAG pipelines — feed fresh, high-signal social text into retrieval systems
• KOL & influencer discovery — profile accounts and their reach in a niche
• Sentiment analysis — classify public opinion on trending crypto/tech topics

Pricing

Pay per result: $0.001 per dataset item (cast or profile).

Example: 10,000 casts ≈ $10. You only pay for results actually delivered; the actor stops when your run's charge limit is reached.

Technical notes & resilience

• Keyless — uses the public Warpcast API (api.warpcast.com) for search + profiles and the public Pinata Hubble hub (hub.pinata.cloud) for channel feeds. No key, no login, no proxy.
• Never crashes — every upstream call retries transient errors (429 / 5xx / network) with exponential backoff and degrades gracefully on a hard block (keeps partial results, exits 0). A bad or empty input still returns output.
• Verified pagination — cursor pagination is confirmed to advance across pages (no duplicate walls).
• Polite pacing — ~3 requests/sec with automatic backoff.
• Charge-limit aware — stops cleanly when your spending limit is hit; no overrun.

Data compliance

This actor accesses only public, logged-out data through Farcaster's public APIs. Farcaster is an open protocol whose casts are publicly broadcast. No login credentials are used. Suitable for research and business use of public data.