TruePPM

Most project tools store dates. TruePPM computes them — a real Critical Path Method engine with Monte Carlo risk analysis, wrapped in a platform where the board card and the Gantt bar are the same object. The engine installs on its own, so you can check the math before you trust anything else here.

Verify the engine (60s)Get Started Try it (0.4 beta)View on GitLab

Start with the engine, not the platform

You should not have to install a platform to find out whether its scheduling is real. TruePPM’s engine is a separate Apache 2.0 Python package — the same code the platform runs on, versioned in lockstep. No server, no database, no account:

pip install trueppm-scheduler

from datetime import date, timedelta
from trueppm_scheduler import schedule, Project, Task, Dependency

result = schedule(Project(
    id="p-1", name="Verify me", start_date=date(2026, 1, 5),
    tasks=[
        Task(id="design", name="Design", duration=timedelta(days=5)),
        Task(id="build",  name="Build",  duration=timedelta(days=10)),
        Task(id="docs",   name="Docs",   duration=timedelta(days=3)),
    ],
    dependencies=[Dependency(predecessor_id="design", successor_id="build")],
))

print(result.critical_path)                                            # ['design', 'build']
print(next(t for t in result.tasks if t.id == "build").early_finish)   # 2026-01-23
print(next(t for t in result.tasks if t.id == "docs").total_float)     # 12 days, 0:00:00

Fifteen working days across two weekends, computed on a Mon–Fri calendar — and docs, off the critical path, carries twelve days of slack before the date moves. Check it against MS Project. That is why the engine ships on its own.

All four dependency types, lead/lag on every link, multi-calendar composition, cycle detection with the offending path, and Monte Carlo with per-task sensitivity. Full engine reference → · View on PyPI →

Want the whole platform, not just the engine? Install with Docker Compose — one command, five services, running in a few minutes — or read What TruePPM’s evaluators check first for a guided tour. Nothing to install yet? Try it covers the zero-config paths.

The Schedule view in Timeline layout on the Atlas Platform Launch sample: task list beside the full-width Gantt chart with the critical path, summary rollups, and the P50/P80/P95 finish-date forecast

Why TruePPM?

Five claims, each one checkable rather than taken on trust — and one page of the ones that go the other way.

Monte Carlo, in the free core

Three-point estimates give you P50/P80/P95 plus a per-task sensitivity ranking — which tasks actually move the finish. Across the commercial field this is a separately licensed product (Acumen Risk, Full Monte, Safran Risk, Primavera Risk Analysis) bolted onto an exported schedule. No other open-source PPM platform ships it at all. The comparison →

The engine is separable

pip install trueppm-scheduler — CPM and Monte Carlo, no server, no account, the same code the platform runs on. No other PPM tool, open or commercial, lets you audit its scheduling engine in isolation.

Board card = Gantt bar

Not synced. Not mirrored. The same row. MS Project has no agile surface; Jira has no critical path; OpenProject keeps board and Gantt as adjacent views. Here, a contributor drags a card to Done and the project manager’s finish date moves — and closing a sprint writes the milestone it feeds a fresh P50/P80 forecast band from measured velocity. The full story →

Nothing that should be free is taxed

Apache 2.0, no seat count, no feature flags, and no telemetry in anything you install. SSO through your own identity provider shipped in the OSS core in 0.4 — OpenProject gates OIDC behind Enterprise, Plane behind Commercial. SSO is not an enterprise feature → (our hosted demo host does ask for an email →)

Computed, not guessed

Every date and forecast is calculated by the engine with a derivation to show for it — which is what makes the schedule safe to hand an AI. An agent can compute, cite, be refused, and reproduce. The read-only MCP server shipped in the 0.4 beta. How grounding works →

…and what it doesn't do

No resource leveling. One constraint type. No cost model until 0.5, no earned value until 0.8. A measured ~2,000-task ceiling in the Schedule view. If you need automatic leveling today, use P6. The full maintained list: What TruePPM doesn’t do yet →

Who is it for?

Persona	Methodology	TruePPM gives them…	Guide
Project Managers	Waterfall / Hybrid	CPM schedules, Monte Carlo confidence intervals, critical path, velocity-suggested Gantt durations, and an automatic milestone forecast band on sprint close (added in 0.3)	Guide →
Scrum Masters	Agile / Hybrid	Sprint board, WIP limits, velocity tracking, burndown, retrospective-to-backlog automation	Guide →
Product Owners	Agile / Hybrid	Backlog management, velocity-based release forecasting, sprint scope protection	Guide →
Team Members	Any	Board-first task view, My Work cross-project surface, real-time updates, offline sync	Guide →
Resource Managers	Any	Per-project capacity preflight, allocation tracking; cross-program leveling in Enterprise	Guide →
PMO Directors	Any	Program coordination, an operational audit log, five-role RBAC, per-project forecasts; portfolio health scores, demand intake, governance in Enterprise	Guide →
Executives	Any	Monte Carlo confidence intervals per project, read-only share links; portfolio RAG status, board-ready reports in Enterprise	Guide →

For everything shipped so far, release by release, see the roadmap.

Next steps

Install

Set up a local instance with Docker Compose in under 5 minutes.

Try the API

Create a project and schedule — tasks, dependencies, and CPM results via curl.

Find your guide

Project managers & waterfall PMs, team members & agile teams, resource managers, executives & PMO directors.

Evaluate

Architecture overview and API reference for technical evaluation.