Features How it works Pricing Blog About Contact
Independent · bootstrapped · 19 people across 7 countries

We built the tool we desperately needed.

In 2023 our founders were an SEO consultant and a stock photographer with the same complaint: every image published on the web ships completely blind. No title, no caption, no keywords, no geotag — and three megabytes of dead weight. So we automated the fix.

0

Images processed since launch

0

Active accounts

0

People on the team

0

Countries, one timezone overlap

Our mission

Every image on the web should be findable

Roughly a third of all Google searches return image results, yet the overwhelming majority of published photos carry zero descriptive metadata. That is not a creative failure — it is a tooling failure. Cameras strip it, exporters drop it, CMSs ignore it.

Our job is to make the correct thing the effortless thing: standards-compliant EXIF, IPTC and XMP written into the file itself, next-gen compression that keeps it all intact, and AI that describes what is actually in the frame — at ten thousand files a batch instead of one at a time.

Not our business

Selling your images. Training models on them. Locking metadata behind our own format.

Very much our business

Open standards, files you own outright, and output that passes every agency validator.

The promise we shipped with

  • Your originals are never used for model training.
  • Everything auto-deletes after 30 days, or instantly on request.
  • EU data residency by default — Frankfurt and Amsterdam.
  • No lock-in: outputs are plain JPEG, WebP and AVIF files.
  • A human answers support tickets, seven days a week.
Still bootstrapped. No outside investors, no growth-at-all-costs roadmap. We are funded entirely by the people who use the product.

Milestones

From a weekend script to 18 million images

The short version of the last three years.

  1. March 2023 The itch

    A 40,000-photo backlog and one ExifTool script

    Youssef, then freelancing on e-commerce SEO, wrote a batch script to push IPTC keywords into a client's product catalog. It ran for nine hours and worked. The client asked if he could sell it.

  2. November 2023 First release

    v0.1 ships to 40 beta photographers

    Upload, write EXIF/IPTC/XMP, download a ZIP. No AI, no WebP, no dashboard. Thirty-one of the forty testers were still using it six weeks later, which was the only signal we needed.

  3. June 2024 The compression engine

    WebP & AVIF conversion with metadata carry-over

    The hard part was never the encoding — it was re-embedding a full IPTC block into a WebP container without corrupting it. Six weeks of work, and an average file-weight drop of 87%.

  4. February 2025 AI metadata

    Vision models start writing the captions

    We moved inference to a modern edge AI engine, cutting per-image analysis to under 400 ms. Titles, alt text and long-tail keyword clusters in 28 languages landed in the same release.

  5. September 2025 Integrations

    WordPress plugin and public REST API

    One API key optimizes an entire WP media library from inside WP-Admin. Agencies pointed it at 30 client sites in a weekend, and the API now carries just over half of all traffic.

  6. 2026 · now What's next

    Shopify app, SOC 2 Type II and structured data export

    Our SOC 2 audit is underway, the Shopify catalog app enters private beta this quarter, and schema.org ImageObject export is on the roadmap for the summer release.

The team

Nineteen people. Four of them talk to you most.

Remote-first from Casablanca to Lisbon, with a standing overlap window every afternoon.

YA

Youssef Amrani

Co-founder & CEO

Ten years of technical SEO for e-commerce. Wrote the original batch script that became the product.

SEO Product
EN

Elise Nordahl

Co-founder & Head of Imaging

Former stock photographer with 60,000 accepted submissions. Owns the metadata standards work.

IPTC Color
RT

Rafael Teixeira

Principal Engineer

Built the edge processing pipeline. Cares unreasonably about p99 latency and encoder benchmarks.

Rust Workers
NH

Nadia Haddad

Head of Customer Success

Answers most of the tickets you'll ever send. Median first reply across 2025: three hours forty minutes.

Support Onboarding

Behind them: 6 engineers, 3 ML researchers, 4 support specialists and 2 designers.

We're hiring — say hello

Values

Four rules we actually argue about

Standards over shortcuts

If ExifTool can't read it, we didn't write it properly. Every field we emit maps to a published IPTC, XMP or EXIF specification — no proprietary sidecars.

Fast is a feature

A batch that takes an afternoon doesn't get run. We budget every release against p99 processing time and refuse work that regresses it.

Your files stay yours

Encrypted at rest, deleted after 30 days, never sampled for training data. The privacy policy is written in sentences, not clauses.

Answer like a person

No ticket deflection funnels, no chatbot maze. Engineers rotate through support so the people who build it hear what breaks.

Come see what proper metadata does

Twenty-five images, no card, about ninety seconds. Then decide.