Sign-in and a redesigned desktop app​

• OIDC single sign-on: sign in with your identity provider (Okta and other OIDC providers) instead of managing static credentials.
• Full GUI redesign on the AWS Cloudscape visual language, upgraded to Angular 22, with a reorganized daemon menu (This Computer / Remote Daemons), redesigned hot-folder cards, and general spacing and icon polish.
• Native OS file and folder pickers on the path fields when connected to a local daemon, so there is no more manual path typing.

Faster many-file uploads​

• AutoMAT (opt-in) automatically scales the number of concurrent transfers to the machine instead of using a fixed cap. Toggle it on in Settings.
• Raised the default max active transfers to 100 and parallelized the "object already exists" pre-check, which significantly speeds up large batches of small files.

Better transfer visibility and control​

• More accurate pause/resume progress: a two-part progress bar separates committed bytes (never moves backward) from in-flight bytes; skipped files are now counted and visible; elapsed time is preserved across a resume.
• Job Details split into peer Files and Logs tabs with sortable columns and status filters.
• Jobs table row actions: a per-row menu mirroring the right-click actions.
• Renamed the confusing "Force transfers" control to "Overwrite existing files".

Remote daemon and hot folders​

• Remote-configuration and S3 bucket-browser UX fixes, plus an auth-lockout fix.
• Hot folders now debounce per folder instead of globally, and a new forceInitialUpload flag controls whether existing files re-upload on start.
• A configured local starting directory is now honored on every connect.

Reliability fixes​

• Fixed two recurring daemon crashes (an event-bus "send on closed channel" panic and a transferstats nil-pointer panic) that could crash the daemon mid-transfer.
• Fixed the bucket report defaulting to the wrong (unauthenticated) configuration.

Distribution and packaging​

• Linux headless CLI daemon binaries are now published (x64 and arm64). Run the engine on a server, render farm, or cloud instance and drive it from a GUI on another machine. See the Headless Linux Installation guide.
• MCP server binaries now ship for macOS (Intel and Apple Silicon), Windows (x64), and Linux (x64 and arm64).
• Windows installers embed a readable version resource, honest OS-tokened asset names, and a hard release gate that fails the build if any expected binary is missing.

Grab the latest build from the GitHub releases page.

High-performance S3 transfers​

• Auto-tuned parallel transfers using native S3 APIs, with no intermediary servers.
• Multipart upload and download with configurable chunk sizes and retry on failure.
• Pause, resume, and cancel active transfers.
• Transfer priority scheduling and job queue management.

Desktop application​

• Native desktop app for macOS (Apple Silicon and Intel) and Windows (x64 and arm64).
• Drag-and-drop file management between the local filesystem and S3.
• Real-time transfer progress, speed, and ETA display.
• File browser for both local and S3 with search and filtering.

Hot folders​

• Monitor local directories for new files and automatically upload them to S3.
• Configurable inclusion/exclusion filters and deduplication.

Checksums and media integrity​

• XXH3, XXHash64, XXHash, and MD5 checksum algorithms.
• MHL (Media Hash List) support for camera-to-cloud verification workflows.
• Post-transfer integrity verification.

Remote daemon architecture​

• Run the transfer engine on any machine (workstation, server, or EC2 instance).
• Control it remotely from the GUI or CLI over an encrypted ConnectRPC connection.
• Transfer profiles for managing multiple daemon connections.

Command-line interface​

• Full CLI for scripting, automation, and headless environments.
• All GUI features accessible from the command line.
• Scriptable job control (start, pause, resume, cancel).

MCP server (fme-mcp)​

• A standalone Model Context Protocol server for AI-assisted file transfers.
• Nine tools: connect, browse local/S3, start upload/download, pause, resume, cancel, and list jobs.
• Supports stdio and streamable-http transports, so AI coding assistants can manage S3 transfers directly.

Cross-platform​

• macOS: Apple Silicon (arm64) and Intel (x64).
• Windows: x64 and arm64.
• Linux: x64 and arm64 (build from source, or use the CLI/MCP binaries).

Download the build for your platform from the 1.0.0 release page.