Etchr

Look like you belong in print. Turn one real photo into a premium editorial portrait pack—recognizable enough to feel true, designed to stand out.

Comment

Etchr — The Editorial Identity Engine

Inspiration

Growing up, I was fascinated by the editorial portraits I saw in publications like The Wall Street Journal.

They were not ordinary photographs. Artists created them by hand through stippling, crosshatching, and thousands of carefully placed lines. Somehow, those portraits could feel more timeless and more personal than the original photograph.

I remember wondering whether I could have one made for myself.

But these were true commissioned works of art, usually created for people being featured in the publication. There was no accessible way for an ordinary person to create one, so the idea stayed with me and eventually faded into the background.

Years later, generative AI changed what was possible.

I came back to the same question:

What if anyone could turn one real photograph into an editorial portrait they would be proud to use everywhere?

LinkedIn.

Company profiles.

Speaker bios.

Personal websites.

Social media.

That became Etchr.

Not a tool for inventing artificial people.

Not a replacement for photographers or artists.

And not another generic AI avatar app.

Etchr is built around a simpler belief:

Everyone deserves one photograph they are genuinely proud to use.

One that still looks like them.

Just a little more refined.

A little more timeless.

A little more intentional.

Technology made Etchr possible.

But technology is not the point.

The point is confidence.

What Etchr does

Etchr transforms one real photograph into a premium editorial portrait.

The experience is built around a simple philosophy:

Recognizable enough to feel true.
Designed enough to stand out.

A user can:

  1. Upload a real photograph.
  2. Confirm image rights and frame the subject.
  3. Create a private Etchr editorial portrait.
  4. Compare the original and finished result.
  5. Save or share the portrait.
  6. Receive a practical portrait pack for profiles, social platforms, editorial use, and personal branding.

The finished portrait is delivered in five useful formats:

Profile

Square

Portrait

Story

Landscape

Etchr is not trying to create a different person. The goal is to preserve the subject’s identity while transforming the photograph into something more intentional and publication-ready.

The hardest product problem

Creating an attractive image was only part of the challenge.

A portrait can be technically impressive and still fail if the person no longer feels recognizable.

We repeatedly evaluated likeness, facial structure, hair, expression, age, and overall character.

We also discovered that even tiny differences in eye position, head scale, mouth position, or crop could make a before-and-after comparison feel artificial.

The viewer did not experience the artistic treatment changing.

They experienced the face moving.

To solve that, we built a canonical facial-registration system that aligns the original photograph and finished portrait inside one stable frame.

For approved comparison pairs, the face now appears stationary while only the artistic treatment changes.

The experience moves from:

“The image shifted.”

to:

“The portrait emerged.”

The system can also refuse to show a comparison when a pair does not meet the visual standard. A strong finished portrait can remain available without being forced into a misleading before-and-after experience.

Founder Review

We did not want to accept an output simply because a model successfully returned an image.

We built a private Founder Review system that separates:

Complete and reviewable

Output not ready

Fixture error

Approve

Revise

Exclude

The system preserves decisions, notes, provenance, request records, checksums, generation costs, and visual-quality evidence.

This allowed us to evaluate the product deliberately instead of selecting only the outputs that happened to look impressive.

It also reinforced an important product principle:

Refusing to show weak work is a feature.

Etchr Labs

During Build Week, we discovered that the underlying transformation generalized far beyond human portraits.

We created Etchr Labs as a private proving ground for a broader idea:

The Editorial Identity Engine.

Using one canonical transformation workflow, we completed a 30-pair benchmark across:

People and identity

Animals and natural detail

Food and drink

Products and objects

Transportation and instruments

Architecture and places

Editorial and narrative illustration

The benchmark included subjects such as:

An owl with detailed feathers

A cheeseburger

A mechanical watch

A vintage field camera

A motorcycle

A lighthouse

Architecture

Packaging

Environmental animal scenes

Twenty-nine of the thirty outputs received founder visual approval.

One was marked for revision because the finished illustration was excellent, but the source and output framing made the comparison slider misleading.

Etchr Labs remains exploratory rather than an unrestricted consumer product promise. It demonstrates that the same underlying system can interpret the visual identity of people, products, animals, food, machines, architecture, and editorial scenes.

Reviewing the collection felt less like testing an image model and more like flipping through an editorial magazine.

How we built it

Etchr was built through a founder-led collaboration with GPT-5.6 and Codex.

Steve — Founder and creative direction

I provided:

The original product idea

The founder vision

The visual references

The product taste

The quality bar

The customer experience

The approval, revision, and exclusion decisions

The business priorities

The final product judgment

GPT-5.6 and Codex did not invent Etchr.

They helped me turn a long-held idea into a real product.

GPT-5.6

GPT-5.6 served as my:

Product strategist

Product manager

Principal product designer

UX and interaction-design lead

Visual-system partner

Architecture and engineering-direction partner

Code-review partner

QA-strategy partner

App Store launch advisor

Build Week storytelling and submission partner

It helped turn subjective founder reactions such as:

“The face is moving.”

“The reveal feels jerky.”

“This does not feel premium.”

into precise product specifications, visual acceptance criteria, and engineering direction.

Codex

Codex served as the implementation and validation agent.

It handled:

Repository verification and recovery

Isolated review branches and worktrees

Implementation of approved specifications

Browser and responsive QA

Playwright automation

Registration measurements and visual overlays

Private Founder Review tooling

Generation ledgers and checksum verification

A 30-pair editorial benchmark

Provenance and cost tracking

Native iOS and Capacitor work

StoreKit integration

Xcode archive and TestFlight preparation

Preview and production validation

Controlled commits, pushes, and draft pull requests

The working model was:

Vision → Design → Approval → Implementation → Validation → Founder Review → Ship

Challenges

Preserving identity

A technically impressive portrait still fails if the person no longer feels recognizable.

We repeatedly evaluated facial structure, expression, hair, age, and overall character.

Facial registration

Small geometric differences made comparisons feel fake.

We audited the full coordinate, crop, scaling, translation, browser-rendering, and CSS-transform chain. We removed compounded legacy transforms and calibrated approved pairs inside one canonical frame.

Designing motion without distracting from the portrait

The reveal initially felt like a wipe, brush effect, or interface transition.

We learned that the portrait must remain spatially immutable and that the final result needs enough stillness to register emotionally.

Honest quality evaluation

Incomplete fixtures originally appeared beside approval controls.

We corrected the review workflow so missing outputs could never be mistaken for failed visual results or accidentally approved.

Privacy and publication boundaries

Founder visual approval does not automatically grant publication approval.

We tracked provenance and checksums, kept benchmark media private, separated visual quality from public-use decisions, and withheld misleading comparisons.

Shipping across web and iOS

The native app loads Etchr’s web application remotely, so production deployment, native bridge behavior, StoreKit, TestFlight builds, and App Store submission all had to be coordinated carefully.

Accomplishments

We are proud that we:

Built a working premium editorial portrait product

Preserved identity instead of producing generic avatars

Created a five-format portrait pack from one result

Solved high-salience facial-registration drift

Built an auditable Founder Review system

Completed a 30-pair Editorial Identity benchmark

Achieved 29 founder-approved benchmark outputs

Tracked provider requests, costs, provenance, and checksums

Prepared Etchr for both the web and iOS

Built the product through a deep founder, GPT-5.6, and Codex collaboration

What we learned

Quality is a system

The generation model matters, but so do framing, registration, review, rights, delivery, motion, and product judgment.

The best technology often disappears

Canonical registration succeeds when the viewer does not notice it.

The person simply remains still while the artistic interpretation changes.

Restraint is a competitive advantage

More motion, more controls, and more features did not automatically make the product better.

Often, the strongest decision was to remove something.

Founder judgment and automation complement each other

Automated measurements, QA, benchmarks, and scoring made review faster and more rigorous.

Founder visual judgment remained the final authority.

The product may be larger than its first use case

Etchr began as an editorial portrait product.

Etchr Labs suggests that the same engine may support products, animals, architecture, food, machines, and editorial storytelling.

What’s next

Near term:

Launch Etchr on the App Store

Complete the StoreKit purchase-to-generation lifecycle

Refine the final processing and reveal experience

Improve gallery and content-publishing workflows

Expand public-safe editorial examples

Continue improving object and scene registration

Make the portrait workflow faster and more reliable

Longer term:

Build Etchr Labs into a curated experimental gallery

Create specialized modes for products, architecture, food, and editorial scenes

Develop a scalable founder Content Studio

Expand the Editorial Identity Engine for individuals, publications, brands, and creative teams

What we believe

People deserve one photograph they are proud to use.

Technology should disappear behind craftsmanship.

Confidence is more valuable than features.

People should still look like themselves.

Restraint is a competitive advantage.

One beautiful artifact is worth more than one hundred average ones.

Software should remove anxiety.

The reveal matters.

Sharing should feel like pride, not promotion.

The portrait comes first.

Everything else is secondary.

Closing

Etchr started with a question I had carried for years:

Why should a timeless editorial identity be reserved for people who appear in the newspaper?

Now one real photograph can become something recognizable, distinctive, and ready to be seen.

Built With

Share this project:

Updates