clipboard_set {image, text, html}: multi-format clipboard write (pup ships its own PowerShell helper today)

by adom

Adom Bridge unleashes your AI onto your full PC: full power, full safety. The AI breaks out of the container it runs in and onto your real machine, managing and running your entire laptop as you: launch apps, move files, drive any window, control KiCad, Fusion 360 or your real signed-in browser. Works with any AI, cloud or local (Claude Desktop, Claude Code, Codex), no Hydrogen required. Bridge and its bridges are an operating system for AI, with managed Node and Python runtimes, and a human-onl

Overview Files Skills Discover Packages Releases Dependencies Issues20 Breadcrumbs Pull Requests1 Activity

Closed general

John Lauer · 2mo ago ·closed by John Lauer

Ask from pup-bridge (Rust port plan, section 3): pup's annotate/snip feature writes the capture to the clipboard as image + text together (a multi-format DataObject: the PNG and a plain-text JSON sidecar, so one paste into any AI chat carries both). pup does it today with a PowerShell STA helper it spawns itself (copyCaptureToClipboard, ensurePupClipHelper).

Please add clipboard_set {image?: path | base64, text?: string, html?: string} writing all given formats in ONE clipboard transaction, and clipboard_get {formats?} for completeness. Then pup (and every other bridge) stops shipping its own clipboard helper.

1 Reply

John Lauer · 2mo ago

Shipped in ab 2.1.33 (insiders), native Win32, no shell-out:

desktop_clipboard_set {text?, html?, image?: <path on the desktop | base64 or data-URI PNG>, returnPrior?} writes every given format in ONE transaction (OpenClipboard -> EmptyClipboard -> CF_UNICODETEXT + CF_HTML (offsets computed) + registered "PNG" + CF_DIB 32bpp -> CloseClipboard). Returns {formats:[...], set, image:{width,height}, prior?}. At least one format required; the PNG is decoded before the clipboard is opened so a bad image never leaves it emptied.
desktop_clipboard_get {formats?: ["text","html","image"]} (default text, back-compat shape kept) returns text, the CF_HTML fragment, and the image as PNG: imageBase64 + image:{width,height,bytes,imagePath} (a file in the screenshot temp dir, janitor-pruned) + available:{text,html,image}.

Live: wrote a JSON text sidecar + an HTML fragment + a 64x40 PNG in one call, read all three back (available all true, image 6097 bytes, dims correct), prior text captured and restored. Same gate as the other input verbs (shell-auto-approve). html/image are Windows today; text everywhere. pup can delete copyCaptureToClipboard / ensurePupClipHelper. Closing.

John Lauer closed this · 2mo ago