Feader

Feader is an AI-native, Web3-friendly RSS reader built for people who follow fast-moving information streams across the open web, crypto networks, and decentralized communities.

Product Direction

Feader combines a focused RSS reading experience with AI-assisted understanding and Web3-aware source discovery.

Core goals:

• Keep RSS subscriptions, unread state, saved articles, and reading flow simple and fast.
• Use AI for summaries, topic clustering, deduplication, article Q&A, tagging, and personalized reading queues.
• Support Web3-native information sources such as DAO forums, governance feeds, Mirror, Paragraph, Farcaster, Lens, ENS-linked identities, and project updates.
• Support non-RSS sources through built-in adapters, declarative extraction rules, and script/plugin integrations.
• Stay friendly to user-owned data, portable subscriptions, and open web standards.

Initial Feature Areas

RSS Reader

• Feed subscription management
• Article aggregation
• Read, unread, saved, and later-reading states
• Search and filtering
• OPML import and export

AI Native

• Article summaries
• Multi-source topic briefs
• Semantic search
• Automatic tags
• Feed and article deduplication
• Personalized daily reading recommendations

Web3 Friendly

• Wallet-aware identity options
• ENS, Farcaster, Lens, Mirror, Paragraph, and DAO source support
• Token-gated or community-specific feed support as a future extension
• User-owned exportable data model

Wallet login uses local Sign-In with Ethereum verification. Set VITE_REOWN_PROJECT_ID to enable the Reown AppKit / WalletConnect QR modal; without it, Feader falls back to an injected EVM wallet when one is available.

Plugin System

• Native RSS and Atom support for standard feeds
• Declarative XPath extraction for simple static HTML or XML sources
• Script-based plugins for complex websites that need custom fetching, parsing, login handling, pagination, or anti-fragile extraction logic
• AI-assisted rule authoring so users can ask Feader to inspect a page and fill XPath selectors for title, link, date, author, content, and next-page fields
• A shared article output contract so RSS, XPath rules, and scripts all feed the same reading pipeline

See docs/plugin-system.md for the initial architecture. See docs/evm-wallet-login-plan.md for the EVM wallet login plan and docs/ai-mcp-cli-plan.md for the AI, MCP, and CLI control plan.

Status

This repository has a Tauri, Rust, Vite, React, and TypeScript baseline. The current implementation includes local-first RSS/Atom source management, XPath source preview diagnostics, EVM wallet login through local SIWE verification, refresh status tracking, normalized articles, read/saved state, a built-in JSON CLI for source/article automation, and a reader workspace backed by SQLite.

CLI

The installed feader executable launches the desktop app when run without arguments. With arguments, it acts as a JSON-first control plane over the same SQLite database and feed adapters:

feader sources list --json
feader source add https://example.com/feed.xml --title Example --category News --json
feader source refresh --all --json
feader source rename 1 "New title" --json
feader source category 1 "Research" --json
feader source delete 1 --yes --json
feader articles list --unread --limit 20 --json

Use --db /path/to/feader.sqlite or FEADER_DB=/path/to/feader.sqlite to target a specific database. Without either, the CLI uses the app data database path for the current platform.