# PAPI: Persistent Adaptive Planning Intelligence [Health: Active]

**Category:** 🤖 Coding Agents  
**Repository:** https://github.com/getpapi/papi  
**GitHub Stars:** 11  
**Views:** 1  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/papi-persistent-adaptive-planning-intelligence

## Description
Adaptive plan/build/review cycles for AI coding assistants, persisted across sessions.

## Tools
Capabilities this server exposes over MCP:

- **orient** — one call returns the current cycle, what's in flight, and the recommended next action. Run it at the start of every session.
- **setup** — scaffold PAPI onto a new project.
- **plan** — break goals into a cycle of right-sized tasks, each with a build handoff your assistant can execute directly.
- **build_list** — list the current cycle's tasks and their handoffs.
- **build_execute** — start a task (creates a branch and handoff) and complete it (records the build report).
- **tutorial_mode** — turn on short explanations and one clear next step in every reply. Handy for your first few cycles.
- **board_view** — read the project board and any task.
- **board_edit** — change a task's status, cycle, priority, or notes.
- **ad_hoc** — record quick work done outside the cycle so it shows in project history.
- **idea** — capture a feature, bug, or research note into the backlog.
- **bug** — file a bug against the board.
- **strategy_review** — step back every few cycles to check direction, not just velocity.
- **strategy_change** — record an Active Decision, with supersede history.
- **zoom_out** — a periodic retrospective across many cycles.
- **join** — join a project a teammate invited you to.

## Claude Desktop Quick Installation
Remote MCP endpoint (confidence: high). Install path detected from listing signals. Add as a URL/SSE server in your client:

```json
"mcpServers": {
  "papi-persistent-adaptive-planning-intelligence": {
    "url": "https://mcp.getpapi.ai/mcp"
  }
}
```

## Documentation

## What PAPI: Persistent Adaptive Planning Intelligence does

PAPI: Persistent Adaptive Planning Intelligence gives AI coding assistants a shared operating record for project work. It organizes development into cycles that move from planning through implementation, review, and release. The assistant can inspect the current cycle, identify unfinished work, create or update tasks, record build results, and advance the project when work is accepted.

The hosted state includes the project board, cycle history, build reports, review outcomes, and Active Decisions. This gives later sessions access to prior work and the reasoning behind important choices. A dashboard at getpapi.ai displays cycles, boards, and decisions, while MCP tools let the assistant maintain that information during normal work.

## How it works

A typical session begins with `orient`, which reports the active cycle, work in progress, and suggested next action. `setup` initializes PAPI for a project. `plan` turns a goal into a cycle of right-sized tasks, with build handoffs intended for direct execution by the assistant.

The build flow uses `build_list` to inspect current tasks and `build_execute` to start or complete one. Starting a task creates a branch and handoff; completing it records a build report. Review tools surface finished builds and record whether work is accepted, needs changes, or is rejected. `release` merges completed work and moves the project into the next cycle.

PAPI: Persistent Adaptive Planning Intelligence also supports work outside the main loop. Board tools can inspect or edit tasks, while `ad_hoc`, `idea`, and `bug` record unplanned work and backlog items. `strategy_review`, `strategy_change`, and `zoom_out` provide periodic direction checks, decision history, and longer-range retrospectives.

## Setup and configuration

The supported connection endpoint is `https://mcp.getpapi.ai/mcp`, using streamable HTTP. The project also publishes `@papi-ai/server` as the supported local runtime. Installation instructions are available for Cursor, VS Code, Windsurf, Codex, Claude Code, and generic MCP clients. Claude Code users can install the PAPI plugin or add the HTTP server directly.

After adding the connection, authentication must be completed through an OAuth browser consent flow. An AI assistant cannot complete that browser step. Until authentication is finished, the server reports that authentication is needed and tool calls do not work. Once connected, run `setup` for the project, followed by `orient` to check its current cycle.

The repository contains installation documentation and the issue tracker; the PAPI engine is hosted and closed source. The repository contents are MIT licensed.

## Tools and capabilities

- Orient a session around the current cycle and next action.
- Scaffold PAPI onto a project.
- Plan cycles and generate build handoffs.
- Start and complete tasks with branches and build reports.
- Review completed work and release accepted work.
- View and edit board tasks, including status, cycle, priority, and notes.
- Record ideas, bugs, and ad hoc work.
- Capture strategic decisions with supersede history.
- Run periodic strategy reviews and multi-cycle retrospectives.
- Manage multiple PAPI projects and project documentation through the documented project and document tools.

## Limitations and notes

PAPI: Persistent Adaptive Planning Intelligence requires OAuth authentication before its tools can be used. It depends on the hosted PAPI engine rather than exposing that engine’s source code in this repository. The service is designed for MCP-capable coding assistants; the repository names Cursor, Claude Code, Windsurf, Codex, VS Code, and DeepSeek Harness installation paths.

_Full upstream README: https://allmcps.com/mcp/papi-persistent-adaptive-planning-intelligence/readme_