inoreader-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| INOREADER_CLIENT_ID | Yes | Your Inoreader API Client ID, obtained from the Inoreader Developer Portal. | |
| INOREADER_CLIENT_SECRET | Yes | Your Inoreader API Client Secret, obtained from the Inoreader Developer Portal. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| prompts | {
"listChanged": true
} |
| resources | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| setup_authA | Authenticate with Inoreader via OAuth 2.0. If no code is provided, returns the authorization URL to visit. If a code is provided, exchanges it once for access tokens (overwrites locally saved credentials; OAuth requests are separate from Zone 1/2 accounting). A consumed code must not be replayed. Read inoreader://usage for workflow and budget guidance. |
| get_unread_countsA | Get unread article counts for all feeds and folders, sorted by count descending. Use this first to understand what needs attention. Costs 1 Zone 1 request. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| get_articlesA | Fetch articles from a feed, folder, tag, or all items. Supports filtering by read/unread/starred status and date range. Costs 1 Zone 1 request per page plus 1 for total_count (disable with include_total=false). total_count is one IDs page of at most 10000; check total_count_is_exact and total_count_continuation before claiming a full total. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| get_article_idsA | Lightweight fetch of article IDs from a stream without full content. Useful for counting or batch operations. For a middle ground with titles, use get_articles with compact=true instead. Costs 1 Zone 1 request. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| search_articlesA | Search for articles by keyword across all feeds. Costs 1 Zone 1 request per page. Uses undocumented but stable search endpoint. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| get_saved_web_pagesA | List saved web pages (manually saved URLs, not from RSS feeds). Supports filtering to find cleanup candidates. Pages can be protected from cleanup by starring or tagging with 'Keep' (via manage_tags add_tag='Keep'). Use filter 'removable' to find pages that are neither starred nor kept -- candidates for review, not authorization to remove. Both count_only and listing apply the filter server-side via the Inoreader API. Set count_only=true to count one filtered IDs page (up to 10000 IDs, with continuation and total_is_exact; not necessarily a collection total) without content (1 Z1 request). Costs 1 Z1 request per page otherwise. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| get_saved_itemsA | Get a bounded union of starred articles, saved web pages, and Keep-tagged items, deduplicated with saved_via membership from the fetched pages. Reads up to max_pages_per_collection (default 1, max 10) pages of 100 per collection: 3-30 Zone 1 requests before read retries. total_count is the fetched union, NOT an account-wide total unless complete=true; continuations and per-collection errors disclose partial coverage. Resume individual streams with get_articles. Preflights the maximum base cost against known remaining quota; unknown quota is not assumed unlimited. See inoreader://usage for budgets, template examples and prompts. |
| get_article_contentA | Get full HTML content for specific articles by ID. Use after get_articles or search_articles to read full content. Costs 1 Zone 1 request. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| batch_edit_subscriptionsA | Deprecated: use categorize_feeds instead (more reliable, folder-centric input). Add multiple feeds to folders. Base cost is 1 Zone 2 request per assignment and 1-3 Zone 1 verification reads; reconciliation may add 1 Zone 2 per shortfall. Preflight requires worst-case headroom; read retries cost extra. Reports only assignments confirmed present on the server. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| list_subscriptionsA | List RSS feed subscriptions with their folders, URLs, and metadata. Supports filtering by folder and searching by title/URL. Returns paginated results (default 100). Costs 1 Zone 1 request. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| suggest_feed_cleanupA | Propose cleaned-up titles for feeds that have no folder assignment. Pure string normalization -- strips vendor boilerplate ("Blog on X", "X's Blog", trailing "RSS"/"Feed", and a trailing site-name suffix). Deterministic: no LLM judgment, same input always yields the same output. Makes NO changes -- apply a proposal with manage_subscription action='edit'. Costs 1 Zone 1 request. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| get_uncategorized_feedsA | Get feeds that have no folder assignment, returned as compact [stream_id, title] tuples. Use this to identify feeds needing categorization, then call categorize_feeds with your assignments. Costs 1 Zone 1 request. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| categorize_feedsA | Assign feeds to folders in bulk. Pass a map of {folder_name: [stream_id, ...]}. Typical workflow: call get_uncategorized_feeds first, decide categories, then call this tool. Base cost is 1 Zone 2 request per assignment and 1-3 Zone 1 verification reads; reconciliation may add 1 Zone 2 per shortfall. Preflight requires worst-case headroom; read retries cost extra. Reads the subscription list back afterwards and retries anything that did not land, so the counts reflect what is actually on the server; sets isError when fewer changes were confirmed than requested. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| reassign_feedsA | Move feeds from one folder to another in bulk. Pass from_folder and a map of {new_folder: [stream_id, ...]}. Base cost is 1 Zone 2 request per assignment (combined add/remove), plus 1-3 Zone 1 verification reads. Reconciliation may add 2 Zone 2 requests per shortfall. Preflight requires worst-case headroom; read retries cost extra. A feed counts as moved only when the subscription list shows it in the destination AND no longer in the source; anything short of that is retried with the add and remove split into separate calls, and reported as unverified with isError set if it still does not land. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| rename_folderA | Rename a folder/label. All feeds in the old folder are moved to the new name. Costs 1 Zone 2 request. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| manage_subscriptionA | Add, edit, or remove an RSS feed subscription. Base cost: subscribe 1 Zone 2 (+1 for optional title/folder edit), edit 1 Zone 2 + 1 Zone 1 read-back, unsubscribe 1 Zone 2. No ambiguous write is replayed; read retries cost extra. An edit returns the feed's resulting title and folders with an |
| manage_tagsA | Mark articles as read/unread/starred, or apply/remove custom tags. To mark as read: add_tag='read'. To star: add_tag='starred'. To unstar: remove_tag='starred'. To save/unsave a web page: add/remove_tag='saved-web-pages'. To protect a saved web page from cleanup: add_tag='Keep'. Supports batch operations on multiple articles. Use friendly names: 'read', 'starred', 'like', 'broadcast', 'saved-web-pages', 'Keep', or any custom label name. Base cost is 1 Zone 2 + 1 Zone 1 per 50 articles; observed shortfalls can require a second write/read round. Preflights both rounds; safe-read retries cost extra. Reports |
| mark_all_readA | Mark all articles in a feed or folder as read. Costs 1 Zone 2 request. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| remove_saved_web_pagesA | Remove saved web pages by article ID. Use get_saved_web_pages with filter 'removable' to find candidates (pages that are neither starred nor tagged 'Keep'). To protect a page from cleanup without starring it, use manage_tags with add_tag='Keep'. Base cost is 1 Zone 2 + 1 Zone 1 per 50 pages, with up to a second round for observed shortfalls (preflighted); reports only removals confirmed on the server. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| list_folders_and_tagsA | List all folders, tags, and labels with their unread counts. Only unread_count is available here; counts including read items require separate paginated get_article_ids calls per tag (1 extra Zone 1 request per IDs page, at most 10000 IDs/page). Costs 2 Zone 1 requests. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| batch_manage_tagsA | Apply different tags to different groups of articles in one call. Useful for triage workflows where you categorize items into multiple groups at once. Base cost is 1 Zone 2 per 50 articles sharing an add/remove pair, plus 1 Zone 1 per 50 article intents to verify, with up to a second round for observed shortfalls (preflighted). Reports verified/unverified per article and per operation, and sets an error when fewer landed than requested. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| analyze_feedsA | Analyze feed health and explicit engagement with a category-pooled Beta-Binomial model (default 12 months, prior strength 10). By default counts starred, liked, broadcast and custom tags, excluding folders and passive system states; engagement_streams selects an explicit positive allowlist. Ordinary read flags are not engagement. Articles are deduplicated across positive streams. Categories with fewer than two engaged feeds use the global prior; multiple actual topic priors are averaged, excluding exclude_topic_folders. Reports exact equal-tailed 90% Beta bounds and posterior shapes only for complete, consistent counts. Missing, capped or failed coverage is unknown, not zero engagement. Retained engagement can exceed retained feed volume: this is flagged without silently flooring the denominator. Historical rates are retention/subscription-affected context, never authorization for automatic actions. Engagement and volume follow continuations within max_requests and the known 25-read reserve. Default sort unengaged_per_month estimates wasted posting volume; skipped sorts ascending by total minus engaged. Sorting engaged_count, title or days_since_newest skips engaged-feed volume reads; volume_sample optionally counts never-engaged active feeds. Costs 3 metadata reads plus engagement pages and optional volume pages EVERY call; no caching. max_requests counts base calls including failed calls; safe-read retries are extra (up to 4 attempts, counted in rate-limit status). Check get_rate_limit_status before large scans. See inoreader://usage for limits and URI examples. |
| get_rate_limit_statusA | Check cached quota observations, reset/cooldown times and conservative local estimates. Costs 0 HTTP/auth requests. Header usage is authoritative only at observation time; missing/expired values mean unknown, not free calls. Every attempted API fetch (including retries and failures) is counted locally; other clients/processes are unseen. See inoreader://usage for request costs, URI examples and all workflow prompts. |
| get_user_infoB | Get current authenticated user information. Costs 1 Zone 1 request. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| extract_classifier_dataA | Pull Inoreader Intelligence summaries for the four verification-tagged streams (read/worth-it, read/not-worth-it, audit/worth-it, audit/not-worth-it) and parse a binary RECOMMENDATION + CONFIDENCE classifier output from each. Optionally also scans a folder/window for the score distribution among unverified articles. The verified set is the labeled dataset for calibration analysis. The population scan is the candidate pool for audit recommendations. Costs 4-80 base Zone 1 requests for the verified set (4 streams, up to 20 pages each), plus up to population_pages if include_population=true; retries are extra. Preflights the maximum page budget plus a 25-request reserve against known quota. Missing streams fail rather than masquerading as complete data; cap flags disclose partial coverage. Reads existing summaries only: never starts paid Intelligence processing. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| analyze_classifier_calibrationA | Build a reliability diagram for the LLM classifier on the verified-recommended slice. Bins articles by CONFIDENCE score, computes per-bin Beta-Binomial posteriors with an empirical-Bayes prior from the marginal worth-it rate, and reports posterior mean + 95% credible interval per bin. Pool-adjacent-violators (isotonic regression, weighted by bin n) gives a monotone-smoothed curve for visualization; per-bin intervals are the source of truth for inference. Reports ECE, monotonicity violations, a threshold diagnostic (lowest-score recommend-read bin gives an upper bound on what's just below threshold by monotonicity), Manski-style FNR bounds (always identifiable), and an audit-conditional recall posterior via Monte Carlo when audit data exists. Optional per-feed and per-folder slices for multicalibration-lite. Calls extract_classifier_data internally; cost is the same as that call. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| recommend_audit_articlesA | Suggest a small batch of recommend-skip articles to read as audits. Scores candidates by bin uncertainty (CI width on the calibration curve), proximity to the decision threshold, and (optionally) sparse audit coverage. Extracts existing summaries once with include_population=true, then computes uncertainty locally (no second API scan). Base cost 4-80 Zone 1 verification pages plus up to population_pages (default 30, max 200); preflights the maximum plus a 25-request reserve. Never starts paid Intelligence processing. Already-audited articles are excluded. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
| find_feed_replacementA | Find a working feed for a subscription whose feed address is broken. Checks where the old address redirects, the feeds the site advertises (), and conventional feed paths, then compares each candidate with the posts Inoreader stored from the old feed. confidence 'high' means the candidate carries posts from the old feed or the old address redirects to it; 'medium' means same site but no shared posts; 'low' is anything else. |
| replace_feedA | Replace a subscription with a different feed URL, keeping its title and folders: subscribe to the new feed, put it in the same folders under the same title, confirm that landed, and only then unsubscribe the old one. If the new feed is already subscribed, its folders are topped up and the old one is removed. Read state of the old feed is not carried over. Use with find_feed_replacement. Costs 2-3 Zone 2 requests and 3 Zone 1 requests. Base costs exclude safe-read retries (up to 4 attempts, each counted); honor cached status/cooldown. Never blindly replay uncertain writes. See inoreader://usage for limits, URI examples and prompts. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| triage-unread | Review unread articles, summarize each, and suggest whether to read, star, or skip. |
| feed-health-review | Analyze feed engagement and identify candidates for unsubscribe or reorganization. |
| organize-uncategorized | Find feeds with no folder, propose cleaned-up titles, and suggest folder assignments based on existing structure. |
| summarize-recent | Digest of recent articles grouped by source with key themes highlighted. |
| analyze-classifier | Build a reliability diagram for the Inoreader Intelligence classifier on verified-tagged articles. Reports per-bin Beta-Binomial posteriors, ECE, monotonicity violations, threshold diagnostic, Manski FNR bounds, and an audit-conditional recall posterior when audit data exists. |
| review-saved-web-pages | Review saved web pages, decide which to keep or remove. Pages protected by starring or the Keep tag are preserved. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| usage | Local guide with URI template examples, all prompt names, paging, quota provenance and safe write rules. Zero requests. |
| rate-limits | Cached/local Zone 1/2 quota observations, UTC local attempt counts, reset/cooldown timestamps and estimate provenance. No HTTP or auth probe: 0 requests. Missing/expired usage is unknown, not free calls. See inoreader://usage for request costs, template examples and prompts. |
| subscriptions | RSS subscriptions with id, title, URL and folders (first 500; total/truncated disclosed). Costs 1 Zone 1 request; read retries extra. See list_subscriptions for local pagination. |
| folders | Folder and tag structure (first 500 each; totals/truncation disclosed). Costs 2 Zone 1 requests; read retries extra. |
| unread-counts | Unread counts per feed/folder (first 500 non-zero entries, descending; total entries/truncation disclosed). Costs 1 Zone 1 request; read retries extra. Counts overlap across feeds/folders: total_unread is not a deduplicated article count. |
TDQS
Scored across 29 tools
Most tools target distinct resources+actions, and descriptions explicitly cross-reference alternatives (get_article_ids vs get_articles compact, batch_edit_subscriptions deprecated in favor of categorize_feeds). However there are several genuinely confusable pairs: get_saved_items vs get_saved_web_pages, mark_all_read vs manage_tags add_tag='read', batch_manage_tags vs manage_tags, and the analytics trio (extract_classifier_data / analyze_classifier_calibration / recommend_audit_articles). The detailed descriptions prevent most misselection, so it's above average but not clean.
Every tool is snake_case and verb-leading (get_, list_, manage_, analyze_, extract_, find_, replace_, rename_, mark_, search_). No camelCase or style mixing; the verb_noun pattern holds throughout except for intentional bulk-prefixed variants like batch_manage_tags.
29 tools is heavy for a single MCP server and pushes past the comfortable range; the set bundles core RSS operations with very niche analytics (calibration reliability diagrams, audit recommendation) that few agents will use. Each tool does have a defensible place, and the deprecated batch_edit_subscriptions adds redundancy, so it's borderline rather than clearly excessive.
Coverage is broad: auth, article reads/search/content, unread counts, folders/tags, read/star tagging, subscription add/edit/remove, bulk categorization and reassignment, saved web pages, feed health, and broken-feed replacement all exist with clear lifecycle paths. Minor gaps: no explicit delete_folder or create_folder operation (folder creation is implicit via categorize_feeds), and article-level deletion is absent, but these are workaroundable.