Burn Mcp Server
Turns your reading inbox into an AI-accessible knowledge base with a time-pressure triage system. Exposes 26 tools for searching, moving, and analyzing bookmarks across four states: Flame (24h to read), Spark (30-day window), Vault (permanent), and Ash (expired). Your agent can search your saved articles, fetch full content from URLs, batch-triage your inbox based on relevance, create curated collections with AI overviews, and auto-watch RSS feeds or social accounts to pull new content into your queue. Works via stdio for local clients or stateless HTTPS for cloud environments. Useful when you want Claude to act as librarian, curator, or research assistant over everything you've bookmarked, with the ability to write structured analysis back into your knowledge base.
- GitHub stars
- 1
- tools
- 26
- auth
- Secret needed
- transport
- STDIO, HTTP
- env var
- 1
- last updated
- Jul 31, 2026
Tools
26 toolsPublic tool metadata for what this MCP can expose to an agent.
search_vaultSearch permanent bookmarks by keyword
Search permanent bookmarks by keyword
No parameter schema in public metadata yet.
list_vaultList Vault bookmarks by category
List Vault bookmarks by category
No parameter schema in public metadata yet.
list_sparksList recently read bookmarks (30-day window)
List recently read bookmarks (30-day window)
No parameter schema in public metadata yet.
search_sparksSearch Sparks by keyword
Search Sparks by keyword
No parameter schema in public metadata yet.
list_flameList inbox — what's about to burn
List inbox — what's about to burn
No parameter schema in public metadata yet.
get_flame_detailFull detail on a Flame bookmark
Full detail on a Flame bookmark
No parameter schema in public metadata yet.
get_bookmarkGet any bookmark by ID
Get any bookmark by ID
No parameter schema in public metadata yet.
get_article_contentGet full article content + analysis
Get full article content + analysis
No parameter schema in public metadata yet.
fetch_contentFetch content from a URL (X, Reddit, YouTube, WeChat)
Fetch content from a URL (X, Reddit, YouTube, WeChat)
No parameter schema in public metadata yet.
list_categoriesAll Vault categories with counts
All Vault categories with counts
No parameter schema in public metadata yet.
get_collectionsList all Collections
List all Collections
No parameter schema in public metadata yet.
get_collection_overviewCollection detail with AI overview
Collection detail with AI overview
No parameter schema in public metadata yet.
move_flame_to_sparkKeep it — worth reading. Optional insight.
Keep it — worth reading. Optional insight.
No parameter schema in public metadata yet.
move_flame_to_ashBurn it. Optional reason.
Burn it. Optional reason.
No parameter schema in public metadata yet.
move_spark_to_vaultPromote to permanent. Optional category.
Promote to permanent. Optional category.
No parameter schema in public metadata yet.
move_spark_to_ashNot valuable enough to keep.
Not valuable enough to keep.
No parameter schema in public metadata yet.
batch_triage_flameTriage up to 20 at once.
Triage up to 20 at once.
No parameter schema in public metadata yet.
create_collectionCreate a topic bundle with initial bookmarks
Create a topic bundle with initial bookmarks
No parameter schema in public metadata yet.
add_to_collectionAdd bookmarks (deduplicates)
Add bookmarks (deduplicates)
No parameter schema in public metadata yet.
remove_from_collectionRemove bookmarks
Remove bookmarks
No parameter schema in public metadata yet.
update_collection_overviewWrite AI overview (theme, synthesis, gaps)
Write AI overview (theme, synthesis, gaps)
No parameter schema in public metadata yet.
write_bookmark_analysisWrite structured analysis back to a bookmark
Write structured analysis back to a bookmark
No parameter schema in public metadata yet.
add_watched_sourceWatch an X user, RSS feed, or YouTube channel. New posts flow into Flame automatically.
Watch an X user, RSS feed, or YouTube channel. New posts flow into Flame automatically.
No parameter schema in public metadata yet.
list_watched_sourcesList all active watched sources
List all active watched sources
No parameter schema in public metadata yet.
remove_watched_sourceStop watching a source
Stop watching a source
No parameter schema in public metadata yet.
scrape_watched_sourcesFetch new content from watched sources on demand
Fetch new content from watched sources on demand
No parameter schema in public metadata yet.
Sponsors
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any languageConfiguration
BURN_MCP_TOKEN*secretLong-lived MCP token from Burn App → Settings → MCP Server
Readme
Burn — Personal Knowledge Base MCP Server
Want this badge in your own README? Grab the markdown at burn451.cloud/developers#badges.
Your reading data as an AI-accessible knowledge base. 26 tools for Claude, Cursor, Windsurf, and any MCP-compatible agent.
Watch the real Claude → Burn walkthrough — the exact read-only prompt, tool calls, sources, and limits are published so you can reproduce it.
Details
- Repository
- github.com/fisher521/burn-mcp-server
- Website
- burn451.cloud
- Package
burn-mcp-server- Transport
- STDIO, HTTP
- Auth
- Required
- Registry
- MCP registry, active