freshrss-feed-reader

FreshRSS Feed Reader

Read-side access to FreshRSS through the Google Reader compatible API. Prefer these tools over raw HTTP — they normalize each item to a flat, transport-safe text + url so downstream steps read the body directly.

When to use

• Triage what is new in the reading-list or a specific feed/label.
• Fetch the full body of specific items.
• Take a delta since a known timestamp (only items newer than a watermark).
• Check unread counts per stream.

When NOT to use

• Subscribe, unsubscribe, label, star, or mark-read → freshrss-subscription-management.
• Push feed items/sources into the knowledge graph → freshrss-knowledge-ingestion.

Prerequisites & environment

Connect via the mcp-client skill against the freshrss-agent MCP server.

Variable	Required	Notes
FRESHRSS_URL	✅	Base URL of the FreshRSS instance
FRESHRSS_USER	✅	GReader account username
FRESHRSS_API_PASSWORD	✅	GReader API password (Settings → Profile → API)
FRESHRSS_TLS_PROFILE / FRESHRSS_TLS_PROFILE_REF	optional	Named outbound TLS policy (verification is mandatory)

MCP_TOOL_MODE (condensed|verbose|both) selects the condensed surface (used below) vs. the one-to-one verbose tools.

Tools & actions

Prefer the condensed tool; it takes action + a params_json JSON string.

Condensed tool	Actions
freshrss_reader	stream_contents, item_contents, unread_count

Key parameters (stream_contents)

• stream_id — user/-/state/com.google/reading-list (all), feed/<feedUrl> (one feed), or user/-/label/<category> (a category). Defaults to the reading-list.
• count — max items (GReader n, default 100).
• order — o oldest-first or n newest-first (GReader r).
• newer_than — unix-seconds watermark; excludes items older than this (GReader ot).
• continuation — cursor from the previous page to resume.

Recipes (params_json)

Newest 25 items across everything:

{"stream_id":"user/-/state/com.google/reading-list","count":25,"order":"n"}

Delta since a watermark (only items newer than a unix timestamp):

{"stream_id":"user/-/state/com.google/reading-list","newer_than":1719792000,"count":200}

Page 2 using the continuation returned by page 1:

{"stream_id":"user/-/state/com.google/reading-list","count":200,"continuation":"<token>"}

Full bodies for specific item ids:

{"item_ids":["tag:google.com,2005:reader/item/00000000abcd"]}