PCP
PCP (Project Context Protocol)
Standardize Context, Not Intelligence.
One Project Context. Every Contributor.
PCP (Project Context Protocol) is an open standard for preserving, organizing, and transferring project context so any contributor—human, AI, or automation—can continue a software project consistently, regardless of who or what takes over next.
Why PCP?
Software projects lose continuity every day.
Knowledge stays inside people's heads.
Important decisions disappear in chat histories.
Documentation becomes outdated.
AI assistants start from scratch because project context is fragmented.
New contributors spend valuable time rediscovering information that already exists.
As projects grow, context becomes scattered across documentation, source code, issue trackers, chats, and individual experience.
PCP exists to preserve project continuity.
🚀 Start Here
📖 New to PCP? Read our practical beginner guides:
Related MCP server: @context-anchor/mcp-server
⚡ Quickstart (CLI & MCP)
1. Initialize Context in Any Repository
npx @craftsolutions/pcp init --name "My Project"2. Verify Context Integrity (Linter)
npx @craftsolutions/pcp check3. Connect to AI Assistants via MCP (Cursor / Claude / Antigravity)
Add this to your IDE's MCP configuration:
{
"mcpServers": {
"pcp": {
"command": "npx",
"args": ["-y", "@craftsolutions/pcp", "mcp"]
}
}
}🔍 Why PCP vs Other Solutions?
Solution | Limitation | PCP Advantage |
README.md | Surface-level landing page; lacks structured decision logs and rules. | 6 mandatory deep-context components. |
Notion / Jira | External to code repository; AI cannot query it while coding. | Lives directly in the Git repository alongside code. |
.cursorrules | Locked to Cursor; useless in Claude Desktop, VS Code, etc. | Universal open standard working across all AI tools. |
MCP | MCP is the pipe (transport). | PCP is the standard context payload flowing through that pipe. |
Repository Guide
Goal | Location |
🇹🇷 Türkçe Tanıtım Rehberi |
|
🇬🇧 Practical Guide |
|
Understand the vision |
|
Understand the philosophy |
|
Learn the principles |
|
Learn the concepts |
|
Read the specification |
|
Accepted RFCs |
|
Official CLI & MCP Server |
|
Copy the official template |
|
Explore a working example |
|
See open work |
|
What is PCP?
PCP defines a shared Project Context standard.
Instead of teaching every AI system how your project works, PCP standardizes the information surrounding the project itself.
This enables:
Humans
AI assistants
Autonomous agents
Development tools
Future contributors
to work from the same understanding of a project.
The Core Idea
Software projects should not depend on memory.
They should depend on shared context.
PCP creates a single, structured Project Context that survives:
contributor changes
AI model changes
team growth
long development pauses
tooling evolution
One project.
One context.
Every contributor.
Core Principles
Context over conversations
Continuity over memory
Concepts over files
Open by design
Human and AI collaboration
Preserve knowledge, don't rediscover it
Standardize context, not tools
How PCP Works
Contributor
│
▼
Project Context
│
▼
Next Contributor
│
▼
Project ContinuityEvery contributor reads from the same Project Context.
Every contributor leaves the project with better context than they found.
What PCP Is
PCP is:
An open standard
A Project Context Protocol
A continuity-first approach to software development
Human and AI friendly
Technology independent
Implementation independent
What PCP Is Not
PCP is not:
an AI framework
an agent framework
a prompt library
a project management tool
a documentation platform
a version control system
a knowledge base application
PCP standardizes context, not intelligence, tools, or workflows.
Reference Implementation
This repository contains the official reference implementation of PCP.
The reference implementation uses Markdown documents to demonstrate the concepts defined by the specification.
Alternative implementations may use different file formats, databases, or storage systems, provided they preserve the required concepts defined by the specification.
The specification defines concepts.
Reference implementations demonstrate practical implementations.
Repository Structure
docs/ Specifications and documentation
rfcs/ Design decisions and proposals
reference/ Official reference implementation
examples/ Example projects
assets/ Images and diagrams
scripts/ Development utilitiesProject Status
⚠️ Active Design
PCP is currently in the specification phase.
The priority is building a stable conceptual foundation before developing tooling and integrations.
Roadmap
Complete PCP Specification v0.1
Publish Reference Context Template
Expand example projects
Build the Compliance Validator
Develop the PCP CLI
Publish VS Code integration
Publish MCP integration
Release PCP v1.0
Vision
We believe software projects should be able to continue regardless of who—or what—takes over.
Project continuity should not depend on memory.
It should depend on a shared, structured, and portable Project Context.
Philosophy
Don't preserve conversations.
Preserve context.
License
MIT License.
This server cannot be installed
Maintenance
Related MCP Servers
- AlicenseBqualityCmaintenanceMCP server for Contextium — pipe team SOPs, coding standards, agents, skills, and workflows directly into Claude, Cursor, and other AI coding assistants.67MIT
- Flicense-qualityDmaintenanceGit-native MCP server for managing AI context across sessions. Enables LLMs to access project and feature context via markdown files, preserving decisions and constraints.1
- Alicense-qualityAmaintenanceCLI + MCP server that bootstraps persistent project context for any AI agent, storing structured information using hierarchical dot notation and exposing it via MCP tools for session-over-session knowledge growth.1MIT
- Flicense-qualityBmaintenancePersistent context layer for AI coding assistants. Sync memory files from Claude Code to ChatGPT, Gemini, Cursor, Windsurf, and Copilot via a native MCP endpoint. Bearer token auth, JSON-RPC 2.0 over stateless HTTP. Free tier available.
Related MCP Connectors
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
One shared context your team's AI tools read & write over MCP. No re-explaining. Free.
Private-by-default, local-first memory/context/task orchestrator for MCP apps and agents.
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- 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/IsoCodeCrafter/PCP'
If you have feedback or need assistance with the MCP directory API, please join our Discord server