The Lair Toolbox

Converts, trims, resizes, and re-encodes video and audio in nearly any format. It burns in captions, mixes audio tracks, and renders final platform-ready files.

Final export step for every client social video, and the tool that burns word-highlight captions onto Felicia's clips before they go out.

Assembles multiple video and audio clips into a single timeline with transitions, titles, and overlays, without opening a GUI editor.

Stitches Higgsfield-generated clips together into one finished video with title cards and transitions, as part of the automated video pipeline.

Trims, filters, and analyzes audio files. Detects silence and adjusts levels.

First pass on raw clip audio, catching dead air before the silence-cutting step.

Transcribes spoken audio into text with word-level timestamps. Runs the small.en model fully on the Lair's own hardware.

Generates the word-by-word timing data behind auto-captioned client social videos. This is our first transcription tool, and because it runs entirely local, no client audio ever leaves the box.

Scans a video or audio file, finds the silent or motionless stretches, and cuts them out automatically.

Tightens up raw Higgsfield clips before captioning, so finished videos don't drag.

Brings audio up or down to a standard, consistent loudness level (EBU R128, the broadcast standard) so volume doesn't jump between clips or platforms.

Final audio pass on client videos so a video doesn't sound quiet on one platform and blown out on another.

Opens and edits SVG vector files, and exports them to PNG, PDF, or EPS. Converts text to outlines so fonts don't break on other machines.

Batch-exporting logo variants for a Brand DAM delivery, in every format and size a client needs at once.

Converts between nearly any image format, resizes, crops, and applies batch edits across hundreds of files at once.

Bulk resizing and format conversion for photo shoots and mockup libraries.

Renders SVG files to PNG or PDF quickly, without the overhead of a full editor.

Fast, scriptable icon and logo exports inside automated build pipelines.

Converts a black-and-white scanned image into clean vector outlines.

Turning a scanned hand-drawn logo sketch into an editable vector file.

Automatically detects the subject of a photo and cuts out the background, no manual masking needed.

Prepping product or headshot photos for a client site or Brand DAM without a Photoshop round-trip.

Strips unnecessary data out of SVG files to shrink their size without changing how they look.

Compressing icon sets and logo files before they ship on a client website, for faster page loads.

Views and edits the hidden metadata embedded in a photo file, things like camera info, copyright, and location tags.

Geotagging client photos so local SEO signals travel with the image file itself.

Each tool shrinks one image format (PNG, JPEG, WebP, or animated GIF) as small as possible with no visible quality loss.

Final compression pass on every image before it goes on a client website, keeping page load times fast.

Lets us write small custom scripts for image tasks that don't have an off-the-shelf tool, like batch analysis or custom filters.

One-off scripting for image QA checks that aren't covered by a dedicated tool.

Generates images, video clips, and other AI media assets from a text prompt, without opening a browser.

Source clip generation for Felicia's social video pipeline, this is where the raw footage starts.

Strips a font file down to only the characters a site actually uses, cutting the file size dramatically.

Shrinking webfont files on Astro client sites for faster page loads.

Converts a full-color photo or graphic into a vector file, unlike potrace which only handles black-and-white.

Converting a flat-color client logo photo into a clean, scalable vector for print and signage.

Compares two screenshots pixel-by-pixel and highlights exactly what changed between them.

Automated visual QA on client site builds, catching layout regressions before a deploy goes live.

Scores how different two images look to a human eye, not just pixel-for-pixel, useful for judging compression or AI-image quality.

Grading AI-generated image variants against each other to pick the best one before a client sees any of them.

Converts between document formats, Markdown, Word, PDF, HTML, and more, in either direction.

Converting a Markdown SOP doc into a clean Word file for a client who needs it outside our system.

Reads, renders, and converts PDF and PostScript (including EPS) files.

Backbone of importing vendor-supplied EPS logo files into our workflow.

Extracts text, images, and metadata from PDF files, or converts pages to images.

Pulling a page out of a client-supplied PDF brand guide as a standalone image.

Runs text recognition (OCR) on a scanned document and adds an invisible, searchable text layer on top of the original scan.

Turning a scanned paper contract or vendor invoice into a searchable, text-selectable PDF.

img2pdf wraps images into a PDF without recompressing them, pikepdf edits PDF structure directly, pdfminer.six pulls text and layout data out of a PDF.

Assembling a set of print-ready images into a single lossless PDF for a print vendor.

Opens and converts Word and Excel files to PDF from the command line, no desktop app needed.

Converting a client's Word or Excel deliverable to PDF for final proof delivery.

Checks a PDF against strict print and archival standards, flagging anything a print vendor would reject.

Preflight check on a print-ready file before it goes to a vendor, catching problems before we pay for a print run.

Runs and builds JavaScript-based websites and tools. pnpm is our standard package manager for client deliverable repos.

Every Astro client site build runs on this stack.

Drives a real browser from a script, for automated testing or taking screenshots of a live page.

Automated QA passes and screenshot verification on client site builds before they go live.

Pushes a built website or Worker up to Cloudflare's hosting.

The deploy step for every client site and for the PixelDrip site itself.

Manages GitHub repos, pull requests, and issues without opening a browser.

Pushing client deliverable repos and managing pull requests during a build.

Tracks every change to a project's files over time, and lets multiple people work on the same codebase safely.

Every client website and internal tool is a Git repository.

Filters and reshapes JSON data from the command line.

Pulling specific fields out of API responses during automation scripting.

Runs custom scripts for automation, data processing, and tool integrations.

Backbone of most of the custom automation scripts on the Lair.

Builds fast, mostly-static websites that ship little to no JavaScript by default.

The standard framework for every new PixelDrip client site build.

Loads a UI/UX design method into build sessions that pushes generated interfaces toward distinctive, production-grade design instead of the generic look AI tools default to. Covers components, layouts, and full pages.

Part of Iris's and Fynn's standard kit on every web and UI build, from QFrame concepts to Command Center screens, so client sites come out looking designed, not generated.

Keeps terminal sessions running in the background, even after disconnecting, and lets multiple windows share one session.

The Lair's main working session stays alive between logins.

Creates a secure private network between devices, so the Lair can be reached remotely without exposing it to the open internet.

Remote access to the Lair from anywhere, without opening public ports.

A simplified front end for Linux's firewall, controlling what network traffic is allowed in or out.

Locks down the Lair's network exposure to only what's needed.

Copies and syncs files between local storage and cloud services (Google Drive, S3, and similar).

Moving client asset backups between the Lair and cloud storage.

Efficiently copies and synchronizes files between locations, only transferring what changed.

Local backups and file transfers between Lair directories.

Shows a live, readable view of CPU, memory, and running processes.

Quick health check when the Lair feels slow or a job is running long.

Shows which folders and files are eating up disk space, in an easy-to-navigate view.

Finding what to clean up when the Lair's disk starts filling.

Runs an application in an isolated, self-contained environment, so it doesn't depend on anything else installed on the Lair.

Runs veraPDF in an isolated, locked-down container so the print-preflight tool can't touch anything else on the box.

Handles PixelDrip-specific recurring jobs, like Lair maintenance tasks, Nova memory pipeline steps, and WordPress fleet operations.

Day-to-day Lair upkeep and the Nova nightly memory cycle.