polyplan-mcp
PolyPlan MCP is a multi-model AI planning orchestrator that runs structured, multi-round planning sessions across multiple AI models/CLI tools, collecting, cross-reviewing, and synthesizing plans into one final master plan.
Core Planning Workflow
init– Set up.plans/and.polyplan/directories for a new sessionRound 1 (
round_1/round_1_context) – Each model independently creates a plan without seeing others'Round 2 (
round_2/round_2_context) – Each model peer-reviews all Round 1 plans and produces a revised master plan (requires ≥2 Round 1 plans)Final Plan (
final_plan/final_plan_context) – One model synthesizes all Round 1 + Round 2 plans into a single implementable master plan (requires ≥1 Round 2 plan)
Session Analysis & Insights
show_status– View which models completed each round, open questions, and conflictsshow_conflicts– Surface disagreements between models (technology choices, approach, assumptions)show_agree– Highlight decisions all models independently agreed on (highest-confidence items)show_questions– List unresolved questions raised by models and whether others answered themshow_diff– Compare a specific model's Round 1 vs. Round 2 plan to see what changedshow_summary– Get a one-paragraph summary of each model's plan (optionally filtered by round)
Session Management & Export
export_plans– Bundle the entire.plans/session into a single shareable markdown fileshow_history– View a full audit log of all actions (model, CLI tool, round, time)clear_plans– Delete plan files by scope (all,round1,round2, orfinal) with a safety confirmation gate
Allows GitHub Copilot (via VS Code) to join structured multi-model planning workflows, contributing plans and reviews alongside other AI coding tools.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@polyplan-mcpStart a multi-model planning session for the new user dashboard feature."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
PolyPlan MCP
A multi-model AI planning MCP server for CLI coding tools
What is PolyPlan?
PolyPlan is a project-local MCP (Model Context Protocol) server that enables developers to run structured, multi-model AI planning sessions across multiple CLI coding tools simultaneously.
When developers use multiple AI models (Claude Code, Copilot, OpenCode, Codex, Cursor, AntiGravity) for planning, PolyPlan provides a structured way to collect, cross-review, and synthesize plans from all models into one final implementable master plan.
How It Works
PolyPlan orchestrates a 3-round planning workflow:
Round 1: Each model creates an independent plan saved to
.plans/without seeing what any other model thinks.Round 2: Each model reads all other models' Round 1 plans (except its own) and creates a revised master plan informed by peer review.
Final: One chosen model (your strongest, e.g., Opus 4.6 or GPT-4o) reads ALL Round 1 + Round 2 plans and produces the final, synthesized, implementable plan.
Installation
npm install -g polyplan-mcpNote: Global install does best-effort user-level MCP registration. Run polyplan-mcp init inside each project to create project-local MCP config and slash-command wrappers.
Quick Start
Initialize in your project
polyplan-mcp initRestart your AI coding CLI/editor
In any connected CLI tool, start Round 1
/polyplan round1 [your problem]Or use the direct command:
/round_1 [your problem]Check status anytime
/polyplan statusOr:
/show_status
All Commands
These are the MCP tool names and the generated slash wrapper names. PolyPlan keeps MCP tools as the execution layer and creates client-specific prompt/command files during polyplan-mcp init where supported.
Slash Command | Description |
| Start Round 1 — create an independent plan (no other models seen). |
| Get the Round 1 prompt — call before generating your plan. |
| Start Round 2 — peer-review all Round 1 plans and write a revised master plan. Requires ≥2 Round 1 plans. |
| Get the Round 2 prompt with all other models' plans injected. |
| Start Final round — synthesize ALL plans into one implementable plan. Requires ≥1 Round 2 plan. |
| Get the Final round prompt with all Round 1 + Round 2 plans injected. |
| Show full session state — which models completed each round. |
| Delete plan files. Specify |
| Show all points where models disagreed in Round 1. |
| Show all open questions raised by any model, and which were answered by others. |
| Show what changed for a specific model between Round 1 and Round 2. |
| Show what ALL models agreed on independently in Round 1. |
| One-paragraph summary of each model's plan. |
| Bundle entire |
| Full audit log — model, CLI tool, round, time, action. |
PolyPlan also generates /polyplan as a routing command where client command files are supported:
/polyplan status
/polyplan round1 [problem]
/polyplan round2
/polyplan finalFile Naming Convention
All plans are stored locally in the .plans/ directory using the following convention:
.plans/
round1-copilot-sonnet4.6.md
round1-claudecode-kimik2.md
round2-opencode-gemini2.5.md
final-antigravity-opus4.6.mdSupported CLI Tools
PolyPlan connects to any tool supporting the Model Context Protocol (MCP):
Claude Code (
.mcp.json,.claude/commands/)Cursor (
.cursor/mcp.json,.cursor/commands/)OpenCode (
opencode.json,.opencode/commands/)VS Code / GitHub Copilot Agent mode (
.vscode/mcp.json,.github/prompts/)GitHub Copilot CLI (
.mcp.json)Codex CLI (global
~/.codex/config.tomlbest-effort postinstall registration)Windsurf / AntiGravity / Gemini CLI workflows (MCP tools and prompts where configured)
See COMPATIBILITY.md for the full client support matrix and limitations.
CLI Tool Specific Usage
Always pass your model name
PolyPlan saves plans using the model name in the filename (e.g., round1-claudecode-sonnet4.6.md). Auto-detection is not always possible, so always pass modelName explicitly:
Call the round_1 tool with modelName='sonnet4.6', problemDescription='...', and plan='...'OpenCode / Gemini
OpenCode with Gemini models requires explicit tool invocation syntax. Use "Call the tool" instead of "Use polyplan":
❌ May not invoke the tool | ✅ Always invokes the tool |
|
|
|
|
|
|
The word "Call" forces direct MCP tool invocation in OpenCode/Gemini instead of a natural-language response.
Claude Code
Claude Code auto-detects as claudecode but does not expose the active model name via MCP. Always pass modelName explicitly:
/round_1 modelName=sonnet4.6 problemDescription="..." plan="..."GitHub Copilot (VS Code)
Works with slash commands directly. Still recommended to pass modelName:
/round_1 with modelName='sonnet4.6' for this problem: ...Why Multi-Model Planning?
Different models have different strengths, blind spots, and reasoning styles. Some excel at architectural structure, while others are better at catching security edge cases. By using 5+ models independently and then cross-reviewing, you catch more issues, resolve conflicts early, and produce a significantly more robust plan than any single model could produce alone.
Contributing
We welcome contributions! Please see our CONTRIBUTING.md for details on how to get started, run tests, and submit Pull Requests.
License
MIT © 2026 PolyPlan Contributors
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
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/IMAFDI/polyplan-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server