Ein MCP-Server für Prowlarr.
cd prowlarr-mcp
uv sync
cp .env.example .envSetze PROWLARR_API_KEY in .env. Der API-Schlüssel ist in Prowlarr unter Einstellungen → Allgemein → Sicherheit verfügbar.
Konfiguration:
Variable | Standard | Beschreibung |
|
| Prowlarr-Basis-URL, einschließlich eines URL-Basispfads |
| erforderlich | Prowlarr-API-Schlüssel |
|
| HTTP-Timeout, größer als 0 und höchstens 300 Sekunden |
|
| Harte Obergrenze, die von |
Related MCP server: MCP *arr Server
uv run prowlarr-mcpBeispiel einer stdio-Konfiguration im Codex/OpenCode-Stil:
{
"command": "uv",
"args": [
"--directory",
"/path/to/prowlarr-mcp",
"run",
"prowlarr-mcp"
]
}search_releases akzeptiert:
query: Textabfrage; eine leere Zeichenfolge fordert aktuelle Releases an.
search_type: search, tvsearch, movie, music oder book.
indexer_ids: optionale Prowlarr-Indexer-IDs.
categories: optionale Newznab-Kategorie-IDs.
limit: Anzahl der Ergebnisse, standardmäßig 20 und durch die Konfiguration begrenzt.
offset: nicht-negativer Ergebnis-Offset.
Die Antwort ist absichtlich kleiner als die vollständige ReleaseResource von Prowlarr, behält aber indexer_id und guid für einen zukünftigen Workflow zur Übermittlung von Downloads. Prowlarr kann limit unabhängig auf mehrere Indexer anwenden, daher erzwingt der MCP-Server es auch für die zusammengefügte Antwort. truncated bedeutet, dass der MCP-Server einen Teil der empfangenen Antwort verworfen hat; es ist kein has_more-Paginationsindikator und kann nicht belegen, ob Prowlarr weitere Ergebnisse hat.
v0.1: Suche nach Releases über die konfigurierten Prowlarr-Indexer.
Verfügbare Indexer und Suchkategorien ermitteln.
Eine ausgewählte Release an einen konfigurierten Download-Client übermitteln.
Prowlarr-Health, Indexerstatus und Suchverlauf einsehen.
Das Repository enthält einen generischen stdio-Client zum Starten des Servers und zum Aufrufen eines Tools:
uv run src/prowlarr_mcp/scripts/mcp_client.py \
--command "uv run prowlarr-mcp" \
--method "search_releases" \
--arguments '{"query": "yani neko"}'Verwenden Sie --cwd, wenn der Serverbefehl aus einem anderen Verzeichnis ausgeführt werden muss, und --timeout, um das Standard-Timeout von 60 Sekunden für Tool-Aufrufe zu überschreiben. Das Skript gibt die strukturierte Tool-Ausgabe als JSON aus und beendet sich bei ungültiger Eingabe, Verbindungsfehlern oder MCP-Tool-Fehlern mit einem von Null verschiedenen Exit-Code.
uv run ruff format --check .
uv run ruff check .
uv run pyrefly check
uv run python -m unittest discover -s testsSearch enabled Prowlarr indexers for Torrent or Usenet releases.
An empty query requests recent releases. Omit indexer_ids and categories to search all enabled indexers and categories. Results preserve Prowlarr's ordering. The indexer_id and guid identify a release for a future download submission capability, which is not available in version 0.1.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | ||
| query | No | ||
| offset | No | ||
| categories | No | ||
| indexer_ids | No | ||
| search_type | No | search |
| Name | Required | Description |
|---|---|---|
| limit | Yes | |
| query | Yes | |
| offset | Yes | |
| releases | Yes | |
| returned | Yes | |
| truncated | Yes | |
| search_type | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The annotations already provide readOnlyHint, idempotentHint, and destructiveHint, so the safety profile is clear. The description adds meaningful behavioral context beyond annotations: results preserve Prowlarr ordering, empty queries mean recent releases, and indexer_id plus guid are not yet usable for downloads. It does not contradict the annotations.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is compact: three short sentences that front-load the main action and then add behavioral clarifications. The future-download-availability caveat is a small but relevant addition that prevents the agent from assuming a download capability that doesn't exist in this version. No sentence is wasted.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the simple parameter set, no required parameters, a complete output schema, and safety annotations, this description covers everything needed to call the tool correctly. An agent can understand the search scope, the empty-query special case, filtering by indexers/categories, and the result-ordering guarantee.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
With 0% schema description coverage, the description compensates by explaining the key query semantics: empty query means recent releases, and omitted indexer_ids/categories search all enabled ones. It does not explicitly describe limit, offset, or search_type, but those are straightforward from the schema defaults and the enum values. This is strong semantic coverage despite a few remaining gaps.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The first sentence states a specific action ('Search'), a specific resource ('enabled Prowlarr indexers'), and the result domain ('Torrent or Usenet releases'). The empty-query and omit-filter explanations further clarify the scope. Although there are no sibling tools to differentiate from, the purpose is unambiguous and complete.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description clearly tells the agent when behaviors change: an empty query requests recent releases, and omitting indexer_ids and categories searches all enabled indexers and categories. There are no sibling tools to compare against, so the lack of alternatives guidance is acceptable; the exclusion of download submission capability is also useful context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
v0.1.0search_releasesScored across 1 tool
There is only one tool, so there is no possibility of confusing it with another. Its purpose is clearly focused on searching releases, making misselection unnecessary.
search_releases follows a clean verb_noun naming convention. With a single tool there is no conflicting pattern, and the name accurately describes the operation.
A single tool is slightly thin for a server named Prowlarr, but it is appropriate for the currently advertised search-only capability. The missing functionality is explicitly reserved for a future version, so the count is reasonable, not excessive.
The search tool offers robust coverage for querying, filtering by categories or indexer_ids, and fetching recent releases. However, there is no way to discover available indexer_ids or continue from a search result to download submission, and the description explicitly notes that capability is unavailable, leaving a meaningful workflow gap.
Search the tool catalog. Customer API key required. Catalog metadata is not a security guarantee.
Search OpenAlex papers, authors, institutions and citations. Hosted on Apify; needs an Apify token. Send your own Apify API token as an Authorization: Bearer header.
Search website motion, specs and recreation packs. Account required; OAuth or bearer API key.
Search a curated directory of agent tools, MCP servers, APIs, protocols and skills. No auth.