Stop re-explaining your
codebase every session.
Every time you open a new session or switch tools, the AI starts from zero. EGC fixes that: one install gives every AI coding tool a shared, persistent memory of your stack, your decisions, and your next steps.
Without EGC -- every session
The Problem
Every AI session starts from zero.
Close the window and the context is gone. Your stack preferences, the architectural decisions you made last week, the approach that failed after three attempts. Next session you spend the first ten minutes re-explaining ground you already covered.
It gets worse when you switch tools. Move from Cursor to Claude Code and you start over again. The AI doesn't know you. It never did.
The Solution
One install. Every tool. Permanent memory.
sh scripts/install.sh detects which AI tools you have and registers the MCP servers in all of them. The AI reads your project state at the start of every session and saves decisions at the end.
Open any session
AI reads your project state and picks up where you left off.
Close any session
AI saves decisions, preferences, and next steps automatically.
Switch any tool
Same state file, same context. No re-explaining.
Ecosystem
Works with every AI coding tool
LLM Providers
Bring your own model
MCP Servers
Two local servers, zero cloud
egc-memory
The one you'll use every session
get_stateReads project memory at session startupdate_stateSaves decisions and next steps, propagates context to all tool config filesstore_decisionPersists a decision to SQLitequery_historyReturns past decisions by timestampsearch_historyFull-text search with BM25 rankingget_project_stateReturns full project state for a given pathworking_memory_setStores transient context with TTLworking_memory_getRetrieves a transient memory entry by keyworking_memory_listLists all active transient memory entrieslesson_saveRecords cross-session knowledge with confidence decaylesson_recallSearches lessons by relevance using FTS5lesson_reinforceIncreases confidence score of a stored lessondetect_patternsSurfaces repeated commands and recurring errorscompress_observationsCompresses raw hook events into typed summaries egc-guardian
Runs in the background
validate_commandBlocks shell injection and unsafe binariesvalidate_writeBlocks writes to sensitive paths (~/.ssh, /etc)reduce_contextDeduplicates and trims Markdown payloadsorchestrate_taskRoutes prompts to relevant agents and skillsauto_learnMines session failures and propagates lessons to all AI tool config files in the project Efficiency
87% fewer tokens to rebuild context
| Metric | Without EGC | With EGC |
|---|---|---|
| Context overhead per session | ~1,500 tokens | ~200 tokens |
| Cost for 20 sessions/month | ~$0.08-0.09 | ~$0.011-0.012 |
| Time re-explaining context | 10 min/session | 0 |
| Shell output entering the model | full command output | crushed up to 99.7% |
The built-in Token Crusher compresses shell output before it ever reaches the model and records every saving in a local ledger. One giant git log went from 545 KB to 1.6 KB in a single crush. Run egc gain to see your own numbers.
Security
Stores decisions and preferences in local SQLite. Never in your source code. Never in the cloud.
Explore
Find us on our channels
EGC is free and open source. If it saves you time and tokens, consider sponsoring. It helps keep the project maintained and moving forward.
Sponsors
Tool Partners
Be the first tool partner
Annual Sponsors
+ Be the first annual sponsorBackers
Monthly sponsors
Be the first
Ready to give your AI a memory?
Install in under 2 minutes. Works with the tools you already use.