Place an Image into a Form Slot
Assigns one image from an uploaded document to one image position registered on one form in the package — available in the appraisal and pre-bid workflows.
Overview
A lot of paperwork has designated places for pictures. A property report has a position for the building sketch and a position for the location map; a photo addendum page takes one picture across its whole area. Each of those is a fixed rectangle on a known page, registered with the form itself — so the call that fills it carries no coordinates, only a name.
The work nobody wants to do is deciding which picture belongs in which position. An inspector comes back with a couple of dozen phone photos in one PDF, and only some belong on the addendum, in a particular order. This tool records one of those decisions: this image, from this uploaded document, on this page of it, goes into this position on this section of the package.
The decision is recorded during the run. Nothing is drawn onto a page at that moment — the images are stamped when the package is assembled, which happens automatically once the run's work is done.
What It Does
| Capability | What it means for you |
|---|---|
| Position-level targeting | The image lands in the rectangle registered for that position, not "somewhere near the top" |
| Page-specific sources | When the source is a multi-page PDF, name the page and that page is what gets placed |
| Section-scoped | Each section carries its own assignments, so a form used more than once in a package addresses its positions separately |
| Fixed geometry | The rectangle and the page come from the form's registered position, never from the call |
| Deferred stamping | The assignment is recorded during the run and stamped when the package is assembled |
When the Agent Chooses This Tool
Two of the three workflows can call this: appraisal packages and pre-bid packages. The coating QA workflow cannot.
In both, the placement is not a judgement call — the workflow's instructions name the placements outright. The appraisal workflow requires one call per comparable-property photo on the comparables page, and says in as many words that none of the three may be skipped. The pre-bid workflow has exactly two positions, one per page of the parsed invitation to bid, and fills them only on the run whose upload they came from.
Where an uploaded filename names its destination, the filename is the position: the rank in a comparable photo's filename is the position's rank, and a subject photo's filename without its extension is the position's name. Where one multi-page PDF feeds a run of sections — a location map, a contractor bid, a photo addendum — the workflow creates one section per page and assigns that page to the section's full-page position.
Two rules constrain what may be assigned. A pre-filled package that a run started from is never assigned to a position, and neither are the per-form documents split out of it; those are sources for text filling only. In the pre-bid workflow the same holds for the invitation to bid itself — only the page images split out of it are placed.
Images that were pulled out of an uploaded document arrive as separate files before the agent runs, already extracted — Extract and Place Images covers that part. Assigning them is this tool's part. And the agent does not go back over its own work: both workflows are told to end the run once the placements and fills have been issued, without polling or verifying.
Inputs and Outputs
Takes: the section being filled, the name of the target position, an uploaded document that is already in the same package, and optionally which page of that document to use.
Returns: confirmation naming the section, the position and the document it now points at. A name the form does not know is refused, and the refusal lists the names that form does accept. The visible result appears when the package is assembled.
Example
An appraisal package needs two things: a photo addendum built from one 24-page PDF of site photos, and a comparables page carrying one photo per comparable property.
For the addendum, the workflow creates one section per page of that upload, and the agent makes one call per section — naming the section, its full-page position, and which page of the upload belongs there. For the comparables page, the three photos arrive as separate image files whose filenames carry their rank, and the rank is the position: three more calls, one per comparable.
Every call records an assignment and draws nothing. When the package is assembled, each named page and each named image is placed into its registered rectangle. The 24-page upload is not modified — it stays available if another position needs a different page.
Limits
- One image, one position, one call. A twelve-photo addendum is twelve calls.
- The position has to be one the form already registers. This tool fills positions that exist; it does not create new ones, and a name the form does not know is refused.
- A position holds one assignment, and a second assignment to a position that is already filled is refused rather than replacing the first. In the appraisal workflow the position can be cleared and then filled again; the pre-bid workflow has no clearing step, so a placement made there stands for that run.
- When the source is a PDF, the named page is rendered whole — nothing extracts the photograph off that page, so a page with margins and a header arrives in the box with its margins and header. When the source is already an image file, naming a page has no effect.
- Images are re-encoded before they are placed. Orientation recorded by the camera is applied and transparency is not preserved. By default the image is fitted inside its rectangle rather than stretched to fill it, though a position can be registered to stretch it instead. Exposure is not corrected — a dark photo stays dark.
- Nothing is visible until the package is assembled. Checking that an image landed correctly means looking at the assembled package, not the source file.
FAQ
Can I see the photo in place before the package is finished?
No. The call records an assignment; the image is stamped when the package is assembled at the end of the run. The assembled package is where you check it.
What happens if a photo ends up in the wrong slot?
A second assignment to a position that is already filled is refused, so the wrong image is not quietly overwritten — the position has to be emptied first. In the appraisal workflow that is what Clear an Image Slot does. The pre-bid workflow has no clearing step.
How does the agent know which photo goes in which slot?
From the workflow's own instructions, which name the placements, and from filenames where those name the destination — the rank in a comparable photo's filename is the position's rank, and a subject photo's filename without its extension is the position's name. Where one multi-page PDF feeds a run of sections, its pages go in order, one per section.
My photos are all in one big PDF — can it pull out just one?
It can name one page of that PDF without splitting the file first. What lands in the position is that whole page rendered as an image, not the photograph cropped off it.
Can the same photo appear in two slots?
Yes. Assignments are held per position and per section, so one uploaded image can be assigned to more than one position. The appraisal workflow does exactly that: the subject's front photo goes on both the subject photo page and the cover.