jackett（jackett 搜索）

Jackett Torznab API

Search configured torrent indexers via Jackett's Torznab API.

Setup

Config (default): ~/.openclaw/credentials/jackett/config.json

{
  "url": "http://localhost:9117",
  "apiKey": "your-jackett-api-key"
}

Quick Reference

List Indexers

# All configured and available indexers
./scripts/jackett-api.sh indexers

# Only configured indexers
./scripts/jackett-api.sh indexers --configured true

Search Releases

# Search across all configured indexers
./scripts/jackett-api.sh search "ubuntu 24.04"

# Search a specific indexer
./scripts/jackett-api.sh search "dune 2024" --indexer nyaasi

# Increase the output window when needed
./scripts/jackett-api.sh search "dune 2024" --indexer nyaasi --limit 50 --offset 0

# Search with a Jackett filter indexer
./scripts/jackett-api.sh search "openwrt" --indexer "tag:public,lang:en"

# Keep raw XML instead of parsed JSON
./scripts/jackett-api.sh search "foundation s02" --raw

Typed Searches

./scripts/jackett-api.sh tvsearch --query "The Last of Us" --season 1 --ep 2
./scripts/jackett-api.sh movie --query "Dune Part Two" --year 2024
./scripts/jackett-api.sh music --artist "Daft Punk" --album "Discovery"
./scripts/jackett-api.sh book --title "The Pragmatic Programmer"

Capabilities

# Inspect search capabilities for an indexer
./scripts/jackett-api.sh caps --indexer nyaasi

# Inspect the aggregate "all" indexer
./scripts/jackett-api.sh caps

Safe Output Defaults

Parsed searches default to --limit 20 to avoid overwhelming the terminal on very large result sets.

• Increase the window with --limit N
• Page forward with --offset N
• Prefer --indexer, --cat, or typed searches such as tvsearch and movie before raising the limit
• Use --raw only when the original XML is required, because it can be much larger than parsed JSON

Search Rules

When searching for releases, normalize the query before calling Jackett:

• Prefer the English title instead of localized titles
• Prefer scene-style format keywords instead of natural-language descriptions
• Keep the query compact; start with title + year/edition + core format terms, then broaden only if needed

Examples:

• Avatar -> avatar
• Avatar: The Way of Water -> avatar the way of water
• Dolby Vision -> dv
• Dolby Atmos -> atmos
• UHD Blu-ray remux -> uhd bluray remux
• Subtitles included usually should not be required in the first query unless the indexer is known to tag it consistently

Recommended search pattern:

• Start with english-title + important format keywords
• Add year when the title is ambiguous
• Add one codec/source token at a time, such as 2160p, web-dl, bluray, remux, x265, dv, hdr, atmos
• If there are no results, remove lower-priority format tokens before changing the title

Response Format

Parsed search results are emitted as JSON with a top-level meta object and results array.

meta includes:

• total, offset, limit, returned, truncated

Each result object includes fields such as:

• title, guid, size, publish_date
• link, details, download_url
• seeders, peers, grabs
• indexer, category
• imdb, tmdb, tvdb

Use --raw to keep Jackett's original XML response.