
Spines
io.github.SuperAgenticPublished by GitHub account SuperAgenticv1.1.0Updated Oct 24, 2026
Access book collections, reading lists, and bookshelf recognition from AI assistants.
Overview
Connects an AI assistant to SPINES book collections, letting it search books, browse shelves, manage reading lists and recognize books from bookshelf photos.
- What you can do
- Search books by title, author or ISBN and read full book details
- Browse collections, categories and personas, and find books commonly shelved together
- Add books to or remove books from reading lists
- Recognize books from one or more bookshelf photos and save them as a collection
- Check the daily recognition credit balance
- Expose collections, books, personas and reading lists as resources, plus prompts for shelf analysis, recommendations and collection comparison
- When to use it
- Useful when you keep a SPINES library and want an assistant to look up books, curate reading lists, or turn photos of a physical bookshelf into a digital collection with analysis and recommendations.
- Requirements
- Runs locally via npm package @spines/mcp-server on Node.js 18 or later, configured in an MCP client such as Claude Desktop, VS Code or Cursor. Requires a SPINES API key in the SPINES_API_KEY environment variable. Optional settings: SPINES_API_URL, SPINES_DEBUG, and SPINES_LLM_PROVIDER with SPINES_LLM_API_KEY. Network access to the SPINES API.
Installation
In SourceWeft
- Open Spines in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.
Other MCP clients
Follow the launch instructions in the repository.
README
SPINES MCP Server
Access SPINES book collections, reading lists, and bookshelf data from AI assistants via the Model Context Protocol.
Features
- Recognize books from photos of a bookshelf; several photos of one shelf become one collection
- Search books by title, author, or ISBN
- Browse and filter book collections
- Manage reading lists (add and remove books)
- Explore personas, categories, and books commonly shelved together
- AI prompts for shelf analysis and book recommendations
Quick Start
Prerequisites
- Node.js 18+
- A SPINES API key (get one here)
Installation
Configuration
Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
VS Code / Cursor
Add to your MCP settings:
Environment Variables
Tools
Read Tools
Write Tools
Tool Details
search_books
get_book_details
get_collection
list_collections
list_categories
get_persona
get_book_pairings
check_credits
add_to_reading_list
remove_from_reading_list
recognize_bookshelf
Send either image_url / image_base64 or images, not both. Example:
- Photo limit: 4 per request with a regular key. A book that appears on two photos is kept once.
- Credits: each photo uses one credit; the request needs at least one to start. With
SPINES_LLM_PROVIDER/SPINES_LLM_API_KEYset, no credits are used. - Prefer URLs. Base64 photos over 4 MB in total do not fit in one request.
- Duplicates: a photo already in SPINES fails the whole request (
RECOGNIZE_IMAGE_DUPLICATE, with the index of that photo). - Timing: the tool waits while the server reads every photo, then polls until book matching finishes (up to 2 minutes). It returns the finished collection, or the job's current status (with its
job_id) if matching is still running or its status could not be read. If the first step times out (5 minutes), the collection may still be created: checklist_collectionsbefore trying again, since a retry uses credits again.
Resources
Prompts
Troubleshooting
Development
License
MIT
Source: README.md at commit 64fb770
Tools
0Version history
3- v1.1.0LatestOct 14, 2026
- v1.0.3Oct 14, 2026
- v1.0.2Sep 16, 2026
Score
50% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
- Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts