The most interesting question in FreeHand’s history is what should survive it.
The installer? The old interface? Every export format? Or the feeling that a group, a path, a page, and a block of text all belong to one manageable piece of work?
Backhand is the project growing out of that question at BonFX. It is an independent, browser-first vector editor built around selected FreeHand workflows, with SVG as the artwork foundation.
Disclosure: this is our project. The description below reflects source, documentation, and test-file inspection on September 17, 2026. It is not a claim of a finished release or a newly completed independent runtime certification.
Start with the behavior people valued
The project uses FreeHand MX as its principal historical reference. Its local research library includes the original manual, page-anchored Markdown, selected figures, errata, and behavior inventories.
That matters because remembered software is an unreliable specification. Two users may remember different versions or operating systems. A command may have had a platform-specific name. A familiar result may have depended on a preference that was set years earlier.
The reference material gives the project a way to ask precise questions: what gets selected, what moves, what remains editable, and what survives saving?
The resulting editor is a new implementation. It is not Adobe’s original source code running in a different wrapper, and it is not an official continuation of FreeHand.
What is implemented in the prototype
The inspected application contains code and interface connections for several substantial workflows:
- Drawing and editing paths, alongside basic shapes and styling.
- Grouping and Paste Inside-style clipping, with access to child objects.
- Multiple pages and page-management controls.
- Layers and layer controls.
- Text blocks and character/paragraph controls.
- Bitmap tracing with preview, cancellation, and editable vector results.
- SVG saving and reopening, plus recovery infrastructure.
- A limited Find & Replace operation covering fill color, stroke color, and stroke width.
There are test files for these areas. Their presence helps explain how the project is being built, but should not be mistaken for a claim that every current combination has just been tested successfully.
Some older documentation describes features as missing even though newer implementation exists. The editorial account therefore follows the code where it is clear, preserves unresolved limits, and avoids a single percentage that would disguise several different definitions of completeness.
Three demonstrations worth following
The first is a clipped composition. Put artwork inside a shape, reach a child, revise it, and retain the relationship. That is more revealing than a screenshot of the menu command.
The second is a small family of pages. Keep related sizes on a shared pasteboard, move between them, and manage the document without treating every variation as an unrelated file.
The third is image tracing. Import a bitmap, preview reduced-color vectors, cancel an unsuitable result, and commit usable paths. The important word is editable: the outcome should join the rest of the drawing rather than become another opaque picture.
These are intended demonstrations of implemented areas. A production claim requires the complete sequence—including undo, save, reopen, and export—to be checked on supported environments.
SVG is a foundation, not a magic compatibility guarantee
An open artwork representation is valuable. It makes it possible to separate much of the drawing from a particular application’s continued existence.
But “SVG-native” does not mean every SVG from every editor will round-trip perfectly. Page arrangements, editor metadata, text behavior, and unsupported constructs need explicit handling.
Backhand’s serializer supports a defined subset and its own document metadata. The right promise is therefore precise: preserve supported artwork and editor state according to a documented contract. Broader interoperability must be demonstrated rather than inferred from the filename.
What remains hard
Typography goes far beyond placing a string on the page. Linked frames, text on paths, font metrics, caret behavior, and stable output all affect real work.
Likewise, page rectangles are not full master-page inheritance. A few graphic replacements are not the whole MX search system. Rendering on screen is not a professional PDF and print pipeline.
Native FreeHand file import is another distinct problem. This article does not present Backhand as a recovery tool for existing .fh archives. For that need, follow the separate file-preservation guide.
Performance also needs honest boundaries. Project reports retain an open condition for a demanding large-image tracing case. A fast small demonstration does not establish unrestricted performance on every machine and document.
The Assistant needs a clear description
The architecture includes a shared command interface and an MCP server implementation. That can make editing operations available to connected tools.
The inspected built-in Assistant, however, currently provides deterministic manual search and a command console. Its source explicitly states that no language model is connected there.
Those are useful foundations. They are not the same as shipping a conversational designer that understands any request. Future AI integration needs its own provider, permission, cost, and behavior decisions.
What this project is not promising yet
There is no launch date, price, public download entitlement, native iPad commitment, or claim of complete FreeHand parity in this announcement.
The first product scope also does not require accounts, collaborative editing, cloud document storage, billing, or the recreation of obsolete Flash delivery infrastructure. Those exclusions help concentrate effort on the drawing and its durability.
This does not mean users who need advanced production features are unimportant. It means the project must state which jobs it can support before asking anyone to depend on it.
A useful way to follow the work
The next meaningful updates should show a complete task and its limits: a drawing reopened intact, a revised clipped group exported correctly, or a text layout that survives another application.
That is the standard laid out in what a FreeHand revival requires. It is also the kind of progress the Backhand development newsletter is intended to cover.
A familiar workspace is a good beginning. The project becomes useful when a designer can finish a job in it and trust the result after the tab is closed.
