AI distillation
Splicr saves the full article and generates a 2-3 sentence insight on top - so your coding agent gets the
quick takeaway without losing the original context.
Example
You save a long RSC article. Splicr stores the full text and adds: "RSC renders on the server, streaming
HTML to the client. No useState in server components. Use 'use client' for interactive parts." Your agent
sees the insight first, full article if it needs more.
Semantic search
Search by meaning, not keywords. Ask "how to handle auth tokens" and find that tweet you saved about
expo-secure-store — even though you never typed those exact words.
Smart routing
Save an article about pgvector. Splicr sees your project uses Postgres + vector search and routes it there.
No folders, no tags, no manual organizing.
MCP native
Add Splicr as an MCP server and it works with Claude Code, Cursor, Codex, or Antigravity. Your saved
knowledge follows you across tools.
Proactive context
Your agent calls get_project_context when
you start a session. Splicr checks what you've saved since it last served context for this project. If
there's new stuff, you get it. If not, it stays quiet.
Example
Monday: you save 3 articles about auth patterns. Tuesday morning: you open Claude Code - Splicr surfaces
those 3 saves. Tuesday afternoon: you open it again - nothing new, Splicr stays silent.
Cursor
Antigravity