Gantt Dependency Planner
Calculate a small project from task durations and predecessor relationships. A forward pass finds earliest dates along the longest dependency paths; a backward pass finds latest dates and total slack against the overall project finish. The visual chart and exact table show the same schedule.
Key features
- Cycle, missing-ID and duplicate-reference validation
- Parallel branches and predecessor delay propagation
- Forward/backward passes with exact integer slack
- Multiple critical links and standalone critical tasks
- Inclusive UTC calendar dates and a labeled Gantt chart
- CSV and JSON downloads from current page memory
How to use
- Load the example or define task IDs and names.
- Set whole-day durations and comma-separated predecessor IDs.
- Choose a project start date and calculate.
- Review dates, slack and every critical link.
- Adjust a duration or relation and recalculate before exporting.
Use cases
- Plan a small assembly with parallel preparation tasks
- Explore how a delayed task changes the finish date
- Learn a deterministic critical-path calculation
Frequently asked questions
Which calendar does the planner use?
Every calendar day, including weekends and holidays. The start and displayed end dates are inclusive. Internally tasks use zero-based start offsets and exclusive finish boundaries. UTC date arithmetic avoids daylight-saving hour shifts; this is not a business-day calendar.
What dependency types are supported?
Finish-to-start with zero lag only: a successor starts on the calendar day after its latest predecessor ends. Durations must be positive integers. Resource leveling, overlap leads, external deadlines and constraints are not modeled.
How are critical paths and slack calculated?
Earliest start is the greatest predecessor finish. Backward from the global project finish, latest finish is the smallest successor latest start; terminal tasks use the global finish. Total slack is latest minus earliest start. All zero-slack tasks and connecting boundary-matched links are marked; multiple critical paths are possible.
Can inputs contain a loop or missing task?
No. Cycles, unknown predecessors, duplicate IDs/references, self-links and empty comma tokens are rejected. No partial result is presented. IDs are case-sensitive. Renaming a task does not silently rewrite user references.
What are the limits and export details?
Use1–60 tasks, durations1–3,650 days, names up to80 Unicode characters and total horizon at most36,600 days. Project start must be in1900–2099; resulting dates may extend beyond2099. CSV includes exact dates, offsets, predecessors and slack. JSON also contains the input and critical edges. Formula-like CSV text is prefixed with an apostrophe.
Does this guarantee a delivery date?
No. It calculates the dependency model you entered, assuming unlimited resources and fixed durations. It does not predict uncertainty, staffing availability or completion probability. Recalculate when assumptions change.
Privacy
Tasks and results remain in page memory and are not sent to this site or analytics events. There is no automatic storage. Export only the files you choose.
Comments & questions