EGC
The EGC server provides persistent AI memory management across coding sessions and tools, allowing AI assistants to remember project context, decisions, and preferences across sessions.
get_state: Retrieve full project memory at session start, including past decisions, preferences, things to avoid, and next steps — scoped per project path.update_state: Save and merge session decisions, preferences, next steps, and avoidances at session end. Non-destructive — merges with existing memory without erasing history.store_decision: Persist a specific decision with its context into SQLite storage, with a lock arbitration mechanism to prevent conflicts.query_history: Retrieve past decisions with pagination support (limit/offset) to review historical choices across sessions.get_project_state: Retrieve the current project state snapshot (alternate/lighter state retrieval).
Key behaviors:
Memory is stored locally (
~/.egc/state/) as Markdown files backed by SQLite.State is scoped per project path, so different projects maintain independent memory.
Designed to be called automatically at session start (
get_state) and end (update_state).
🌐 English · العربية · Español · हिन्दी · 日本語 · 한국어 · Português (Brasil) · Русский
EGC - Extended Global Context
Your AI agents never start from zero again.
Zero setup. Zero commands. You work, EGC remembers.
EGC is a local runtime that gives every AI coding tool you use a persistent memory. At the end of each session, your AI saves what it learned: decisions made, what failed, your preferences, what to pick up next. At the start of the next session, it loads that state back on its own, no prompting required. Say "let's continue" or "where did we stop?" in any language and your AI already knows what to do. One install covers Claude Code, Cursor, Gemini CLI, Windsurf, Zed, VS Code with GitHub Copilot, and more. Works with Claude, GPT-4o, Gemini, and OpenRouter models including DeepSeek, Qwen3, and Llama 4.
Your AI already knows
You open Claude Code on a project you haven't touched in two weeks. Without typing anything:
State loaded from egc-memory via ~/.egc/state/MyApp/main.md
Context and preferences acknowledged.
Ready to pick up:
• Fix the rate limiter edge case on concurrent requests
• Add integration tests for the new auth module
• Review open PR from @contributor before merging
=== EGC Stack Briefing ===
Stack: typescript, node
Skills: tdd-workflow, coding-standards
Agents: code-reviewer
Guardian: active, every command checked before it runs
===This isn't a cache of your last chat. EGC remembers the decisions, the dead ends, and your preferences, and it stands guard the whole session, blocking the commands that would burn your codebase down before they run. You didn't ask for any of it. You just started working.
Related MCP server: Tages
Install
Same install command on Windows, macOS, and Linux:
npm install -g @egchq/egc && egc installWindows has a few of its own caveats (PowerShell version, Antigravity CLI, Gemini CLI's discontinued free tier): see the Windows notes if you hit anything unexpected.
Or run without installing globally:
npx @egchq/egc installOne brain, many tools. With the GitHub Copilot Chat extension installed, Copilot finds the skills on its own, and the same memory you already have in Claude Code or Cursor shows up there too:
npm install -g @egchq/egc
egc install --target copilotWhat EGC gives your AI
EGC always runs two things together, every session: a memory that keeps what matters, and a safety layer that blocks dangerous commands before they run. It all comes ready, no configuration needed.
Memory: what your AI remembers on its own
You'll never memorize a single command. Say it in any language: "continue from yesterday", "remember this decision", "what broke last time", and your AI knows exactly what to do. The work is yours, the remembering is EGC's.
egc-memory
Tool | What it does |
| Loads everything your AI already knew about the project the moment the session opens |
| Saves what got decided today so nobody loses the thread tomorrow |
| Writes down one important decision, for good |
| Shows past decisions in the order they happened |
| Finds anything that was ever decided, even if you don't remember the date |
| Quick notes that expire on their own once they stop being useful |
| Records something learned, which fades over time if nobody confirms it again |
| Brings back the lessons that are still worth acting on |
| Reinforces a lesson when it gets confirmed again |
| Notices when the same error or command keeps repeating |
| Summarizes the raw history so you don't burn tokens for nothing |
| Checks that memory is working the way it should |
Every branch of your project keeps its own memory, encrypted on your machine: nobody else has access, not even the cloud. Privacy by default, nothing to configure.
Context and safety: what stands guard while you work
egc-guardian
These tools run automatically in the background. Every shell command and every file write is checked before it executes. You never invoke them directly.
Tool | What it does |
| Checks every command before it runs: blocks the ones that could cause damage |
| Stops the AI from writing to sensitive files by accident |
| Shrinks large files so you don't burn your token budget for nothing |
| Picks the right tools for each request, without you needing to know which ones exist |
| Learns from the session's mistakes and writes it down so it doesn't repeat |
Enforced, not requested
Security that doesn't depend on the AI being in a good mood: every command passes through EGC before it runs, always. Full details on harness enforcement, session-intent detection, and the memory miner →
One memory. Every tool you use.
Run egc watch once and forget it exists. Change context in Cursor, it shows up on its own in Gemini CLI, Copilot, Windsurf, Zed: everywhere you work. No manual steps, no stale state anywhere.
egc watch # watch current project
egc watch /path/proj # watch a specific project
egc watch --quiet # suppress outputDashboard: watch your agents work
See every tool call, token, and cost your agents generate, live in your browser. Starts automatically after egc init. Full guide
Prompt library
As a bonus, EGC gives you access to 63 agents, 230 skills, and 77 commands, plus 111 rules: specialists that review your code on their own, best-practice guides for every language and situation, shortcuts that run a whole sequence of tasks for you, and style rules that keep your code consistent. All written from real engineering sessions, not theory. Don't want to use any of it? Fine: EGC's persistent memory works exactly the same.
🌐 English · العربية · Español · हिन्दी · 日本語 · 한국어 · Português (Brasil) · Русский
Support EGC
EGC is built by one developer, maintained in the open, and free.
Website: full docs, feature overview, and live demo
Join the Discord: ask questions, share feedback
Sponsor on GitHub: any amount
Donate via PayPal: no GitHub account needed
Star the repository: helps other developers find it
Contribute: agents, skills, commands, bug fixes, docs
Share: if EGC changed how you work, tell someone
Sponsors
Support from the community keeps this project alive and independent.
Tool Partners
AI coding tools that integrate natively with EGC. Partners get logo placement across all READMEs and EGCSite.
Annual Sponsors · Be the first annual sponsor.
Backers
Monthly sponsors · be the first
Maintenance
Latest Blog Posts
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Fmarzochi/EGC'
If you have feedback or need assistance with the MCP directory API, please join our Discord server