GIFM v0.5.2 is a local GIF maker and compressor for Discord-ready animated GIFs. It converts MP4, MOV, WebM, AVI, MKV, and existing GIF files with bundled FFmpeg, then retries width, FPS, and palette settings until the output fits the selected target.

• Discord Free 10 MB, Nitro Basic 50 MB, Full Nitro 500 MB, Emoji 256 KB, icon/avatar, and custom target presets.
• Video-to-GIF conversion from common video files.
• Existing GIF recompression through the same palette workflow.
• Output as animated GIF, APNG (Discord stickers), animated WebP (Discord-native, much smaller than GIF), muted MP4 (Discord autoplays inline), or animated AVIF (max compression).
• Multi-file batch submission through the same local queue and current preset.
• Timeline editor for long videos with exact timecodes, playhead start/end marking, saved GIF cuts, and export-one/export-all actions.
• Prepared source sessions so one long video can be staged once and reused for many GIF cuts without re-uploading it each time.
• Visual trim timeline with thumbnail filmstrip context, source duration, resolution, FPS, codec, and rotation metadata.
• Browser-side metadata/frame preflight for supported files, with automatic FFprobe fallback when local preview decode is unavailable.
• Auto-fit loop for width, frame rate, color count, duplicate-frame removal, nth-frame dropping, transparency rectangles, and optional duration trimming.
• Local preview, source/output comparison, Discord color-crush simulation, motion hiding, queued/running progress, cancellable jobs, FFmpeg log, exact output byte count, and download/open-output actions.
• Output metadata cards for dimensions, duration, FPS, frame count, SSIM quality, and Discord readiness checks.
• Output suitability and attempt strategy copy that says whether the GIF fits the selected Discord target and which compression lever was used.
• Persisted settings, named presets, and recent outputs in browser storage.
• Dark-first premium workspace with compact runtime trust cues, calmer settings hierarchy, refined empty/error states, and advanced transforms grouped behind a disclosure surface.
• Language picker with English, Spanish, French, German, and Japanese interface chrome.
• Keyboard-visible focus states, reduced-motion support, ARIA progress, and an output alt-text helper.
• Diagnostics panel with FFmpeg/FFprobe versions, platform info, source estimate, latest FFmpeg command, and copy/download JSON bundle.
• Save-as output flow using the browser file picker when available, with download fallback.
• Optional user-provided gifski backend with separate quality and motion-quality controls while keeping GIFM's bundled FFmpeg path as the default.

npm install
npm run dev

Open the Vite URL shown in the terminal. The API runs on http://127.0.0.1:4174.

npm run build
npm run preview

npm run preview serves the built app and API from http://127.0.0.1:4174.

npm run package:portable
npm run package:smoke

The portable artifact is written to release/GIFM-v<version>-win-x64/ and zipped beside it. It includes a desktop GIFM.exe, the built client, Express server, current Node runtime, production-only node_modules (devDependencies and non-Windows FFprobe binaries are pruned), bundled FFmpeg/FFprobe, caption font assets, the Microsoft Edge WebView2 bootstrapper, and a start-gifm.cmd compatibility wrapper. Launch GIFM.exe to open GIFM as a Windows desktop app with its local processing service managed in the background. To update a portable copy, replace the folder with a newly generated package.

GIFM's desktop shell requires the Microsoft Edge WebView2 Runtime, which ships with current Windows 11. If it is missing, GIFM.exe runs the bundled MicrosoftEdgeWebview2Setup.exe bootstrapper on first launch to install it (this needs an internet connection once); otherwise install the runtime from Microsoft and relaunch.

npm run typecheck
npm run build
npm run test:unit
npm run test:smoke
npm run test:ui
npm run package:portable
npm run package:smoke

The unit test suite (npm run test:unit) covers the pure encoding-strategy logic in server/encoding.js: auto-fit attempt stepping, settings parsing/clamping, target presets, square dimension locks, and protected-path retention matching.

The smoke test generates a small local MP4, uploads it to GIFM, waits for the job to finish, downloads the result, validates the GIF header, and checks that the file fits the configured byte target. The UI smoke test serves the built app and verifies the default English interface renders through the shared string catalog.

Caption overlays render with the Anton typeface bundled under the SIL Open Font License 1.1 at assets/fonts/ (license text in assets/fonts/Anton-OFL.txt).

Convert without the UI using the same local engine:

npm run cli -- input.mp4 --target free --format gif --width 480 --out .
npm run cli -- --watch .\incoming --target nitro-basic --format webp --out .\gifs

The CLI supports the full settings surface: --target, --format, --width, --fps, --duration, --start, --colors, --dither, --bayer-scale, --palette-mode, --per-frame-palette, --loop, --speed, --playback, --rotate, --flip-h, --flip-v, --color-filter, --saturation, --crop-x/y/w/h, --caption-top/bottom, --encoder, --gifski-quality, --no-optimize, --no-auto-fit, --allow-trim, and --preset <file.json>. Run npm run cli -- --help for the full list. The CLI starts the local server, submits the job, and writes the result to the output directory. --watch <folder> auto-converts new video/GIF files dropped into the folder, waiting for stable file size before processing.

Generated GIFs are stored under data/output/. Uploaded sources, smoke artifacts, and temporary work files stay under data/; the directory is ignored by Git.

GIFM does not bundle gifski. To enable it, install a gifski binary yourself and point GIFM at it before starting the app:

$env:GIFM_GIFSKI_PATH = "C:\Tools\gifski.exe"
npm run dev

The Encoder setting then exposes gifski beside the default FFmpeg palette encoder. gifski is AGPL-licensed unless you use a commercial license, so verify the license before redistributing any package that includes or depends on it.

GIFM does not bundle gifsicle either. When a gifsicle binary is on PATH (or GIFM_GIFSICLE_PATH points at one), the "Optimize with gifsicle" setting runs a gifsicle -O3 pass on each encode and uses lossy LZW compression (--lossy) as an additional auto-fit lever, which significantly improves the chance of hitting tight targets such as the 256 KB emoji ceiling. gifsicle runs as a separate process and is GPL-2.0-licensed; verify the license before redistributing a package that depends on it.

$env:GIFM_GIFSICLE_PATH = "C:\Tools\gifsicle.exe"
npm run dev

The portable package includes opt-in register-shell.ps1 and unregister-shell.ps1. Run register-shell.ps1 from inside the portable folder (no administrator rights needed) to add a "Make GIF with GIFM" entry to the right-click menu for common video and GIF files; selecting it launches GIFM with that file staged for trimming and export. Run unregister-shell.ps1 to remove it.

Paste a video URL into the import field to download it with yt-dlp. GIFM does not bundle yt-dlp; install it on PATH or point GIFM_YTDLP_PATH at the binary. Downloads are capped at the upload limit and staged as a prepared source for trimming and export.

$env:GIFM_YTDLP_PATH = "C:\Tools\yt-dlp.exe"
npm run dev

GIFM binds to 127.0.0.1 by default and rejects non-local hosts unless GIFM_ALLOW_REMOTE=1 is set on a trusted network. Uploads are limited to 20 GB by default and are checked before FFmpeg runs.

Optional environment controls:

$env:GIFM_MAX_UPLOAD_MB = "20480"
$env:GIFM_DATA_MAX_MB = "25600"
$env:GIFM_DATA_MAX_AGE_HOURS = "24"
$env:GIFM_MAX_CONCURRENT_JOBS = "1"
$env:GIFM_OUTPUT_DIR = "D:\GIFM-output"
$env:GIFM_GIFSKI_PATH = "C:\Tools\gifski.exe"
npm run dev

Completed outputs, uploads, and temporary work files are pruned by age and total size so abandoned runs do not keep filling disk.