Advanced

Importing from Image

Farline AI can extract a project plan from a screenshot, exported Gantt chart, or any visual project timeline. Upload an image and AI vision will identify workstreams, work items, sizes, and dependencies to create a ready-to-forecast project.

What gets imported

When you import from an image, Farline AI uses AI vision to analyse the visual and creates a new project with:

  • Project name: inferred from any visible title or heading
  • Scenario: Baseline Scenario
  • Workstreams: identified from visual groupings, swimlanes, or color-coded sections
  • Work items: extracted from task bars, rows, or labeled elements with inferred sizes
  • Dependencies: detected from arrows, lines, or sequential positioning
  • Milestones: identified from diamond markers or labeled dates
  • Start date: read from the timeline axis or inferred from visible dates

Supported formats

FormatExtensions
PNG.png
JPEG.jpg, .jpeg
WebP.webp
PDF.pdf (first page)

Maximum file size: 10 MB

What works well

Import from Image works best with:

  • Gantt charts from tools like MS Project, Smartsheet, Monday.com, or Jira Advanced Roadmaps
  • Spreadsheet-based project plans with clear columns (task name, duration, dependencies)
  • Screenshots of timeline views with visible task bars and dates
  • Hand-drawn project timelines with legible text

Example: Before and after

Source image

Upload a Gantt chart or project timeline screenshot like this:

Sample Gantt chart to import

Result in Farline AI

After import, Farline AI generates the project definition and displays the forecast as an Excalidraw chart:

Excalidraw chart generated from imported image

How to import (step-by-step)

From the chat + menu

  1. Open the Farline AI app.
  2. Click the + button to the left of the chat input.

Plus button import menu

  1. Select Import from Image.
  2. The image import dialog opens. Either:
    • Drag and drop your image onto the upload zone, or
    • Click Browse Files to select from your computer

Image import dialog

  1. Farline AI analyses the image with AI vision. This typically takes 10-20 seconds.
  2. On success, the extracted project definition loads into the editor and the forecast chart appears.

From the top menu

You can also access Import from Image via Menu > Import > From Image (coming in a future update).

Tips for best results

  • Use high-resolution images - Larger, clearer images produce better extraction results
  • Include visible dates - Charts with a date axis give the AI better context for scheduling
  • Show task names clearly - Ensure task/work item labels are legible and not cut off
  • Include dependency arrows - Visual dependency lines help the AI map relationships correctly
  • Crop to the chart area - Remove toolbars, sidebars, and other UI chrome for cleaner extraction

After importing

  1. Review the definition - Open the Advanced Editor tab to check the extracted workstreams, work items, and dependencies
  2. Adjust sizes - AI-estimated sizes may need refinement based on your team's estimation approach
  3. Verify dependencies - Check that dependency chains match the source chart
  4. Run a forecast - The forecast generates automatically once the definition is valid
  5. Compare to source - Use the Excalidraw chart to visually compare against your original image

Troubleshooting

Import produces few or no work items

  • The image may be too small or low resolution. Try a higher-quality export.
  • Complex charts with overlapping elements can confuse extraction. Crop to a cleaner section.
  • Charts with very small text may not be readable. Zoom in and re-screenshot.

Dependencies are missing or wrong

  • If the source chart doesn't show explicit dependency arrows, the AI may not infer them. Add dependencies manually in the editor.
  • Sequential positioning (task A ends where task B starts) is sometimes interpreted as a dependency, sometimes not.

Import fails with an error

  • Check that your file is under 10 MB.
  • Ensure the file is a supported format (PNG, JPEG, WebP, or PDF).
  • PDFs are processed first-page only. If your chart spans multiple pages, screenshot the relevant page instead.

Extracted dates look wrong

  • The AI reads dates from the chart's timeline axis. If dates are partially obscured or in an unusual format, they may be misread.
  • You can correct start dates in the Advanced Editor after import.

Last updated: 2026-03-09