observeLABSCO SUMMARY
Keeping everything and spending less context are not opposites here.
The usual tradeoff is forget-or-pay; compressing on the way in and retrieving selectively is what breaks it. Two things distinguish it from a notes plugin: explain_decision reconstructs why something was chosen by walking the evidence chain, and the file-claiming tools exist because parallel agents step on each other. Everything stays local — no cloud tier, no premium tools.
What it is
context-mem ingests every tool call, summarises it with a content-aware summariser, and writes it into a navigable markdown vault backed by SQLite. The store is searchable through MCP with hybrid BM25 plus vector search and an optional LLM judge, so a new session starts with what the last one learned.
What you get
- Core memory —
observestores with auto-summarisation, importance scoring and entity extraction;recallreturns verbatim content by filter;searchruns hybrid retrieval;askanswers in natural language over the whole store;timelineandstatsshow what accumulated and what it cost - Knowledge graph —
save_knowledgewith contradiction detection and validity windows,search_knowledge,promote_knowledgeto a global cross-project store,global_search,resolve_contradiction,graph_query,add_relationship,graph_neighbors - Temporal and intelligence —
temporal_queryfor what was true at a point in time,time_travelto compare two timestamps,explain_decisionto walk the evidence chain backwards,predict_loss,generate_storyfor a PR description, standup update or ADR - Sessions and agents —
wake_upfor a token-budgeted primer,restore_session,handoff_session, plusagent_register,agent_status,claim_filesandagent_broadcastfor parallel agents - 14 summarisers apply different compression by content type; a session of 50 tool outputs went from 365 KB to 3.2 KB, a verified 99.1% saving, with recent high-importance observations kept verbatim and pinned entries never compressed
- The vault at
.context-mem/vault/is plain markdown — open it in Obsidian or grep it; the SQLite store is the authoritative record
Requirements
Node.js. npm i context-mem && npx context-mem init detects your editor and writes the MCP config — Claude Code needs no restart, Cursor, Windsurf and VS Code do. Manual setup registers npx context-mem serve. For Claude Code, init also writes 8 hooks into .claude/settings.json that inject relevant memories on prompt submission, so you do not call observe by hand. The vault is opt-in in v3.4 via .context-mem/config.json and default-on in v4.0. A local dashboard runs at http://localhost:3141. Registry version 4.0.0. MIT.
Setup effort
One command — npm i context-mem && npx context-mem init
USE CASES
Starting a session already knowing the project
Pull a token-budgeted primer of what matters before the first question instead of re-explaining context.
Reconstructing why a decision was made
Walk the evidence chain backwards to the observations that led to it, with dates attached.
Writing the PR description from what actually happened
Generate the summary from the stored session rather than from memory of it.
Running two agents on one repo without collisions
Register each agent, claim the files it is editing, and broadcast findings to the others.
41 tools
Listed in project docs. Your client may show a different list.
recallsearchasktimelinestatssave_knowledgesearch_knowledgepromote_knowledgeglobal_searchresolve_contradictionmerge_suggestionsgraph_queryadd_relationshipgraph_neighborstemporal_querytime_travelexplain_decisionpredict_lossgenerate_storyentity_detectfind_tunnelswake_uprestore_sessionhandoff_sessionagent_registeragent_statusclaim_filesagent_broadcastconfigurebudget_statusbudget_configuresummarizeexecuteindex_contentsearch_contentlist_peoplelist_topicsimport_conversationsbrowsediagnosticsBEFORE YOU INSTALL
What to know first
Tools you get
41 tools named in the project's documentation.
Requires
npx on PATH
INSTALLATION
Connect your client
Server setup
npm i context-mem && npx context-mem init