EV1 Labs
← EV1 Labs
Selected work · verified 21 Jul 2026

Four products. Four different hard problems.

The primary commercial selection stays ordered by evidence, not novelty: public App Store release, external beta submission, native workflow beta and an on-device product prototype.

Discuss a similar problemSee services
PublicInciary 1.0.11 is live under EVL Labs.
BetaHunt of Secrets has a valid build submitted for external beta review.
BuildSamataPro and GymPilot are shown at their real maturity.
LabFour Build Week projects are presented as a separate experimental collection.
Public proof

Choose the closest operating problem.

Open the detailed page for the challenge, current signal, what became clearer and what should not be inferred.

Inciary native iPhone dashboard with demo dataApp Store · 1.0.11Inciary

Professional vertical product: native workflow, privacy, localization, release and paid-plan foundations.

 Hunt of Secrets native iPhone game creatorExternal beta submittedHunt of Secrets

Consumer AI product: real-space game creation, camera matching, backend tests and store package.

 SamataPro native service intakeNative betaSamataPro

B2B workflow: intake, estimating, LiDAR evidence, marketplace roles and human review.

 GymPilot native beginner workout autopilotOn-device prototypeGymPilot

Product simplification: beginner training reduced to one safe action, timing and progression logic.
OpenAI Build Week 2026 · separate collection

Four ways to use AI without handing it the final word.

These competition projects explore release truth, agent coordination, SaaS portability and human access. Every page states what GPT-5.6 may do, who retains authority and what the public demo does not prove.

ProofLatch release decision desk showing a deterministic blocked verdictRelease evidenceProofLatchSubmitted

Deterministic policy owns BLOCKED or READY. GPT-5.6 explains the evidence and prepares a bounded Codex repair brief.

Rules decide · model explainsView case → Gatehold replay showing a semantic workstream holdAgent operationsGateholdSubmitted

A deterministic local clearance layer coordinates overlapping agent work, machine capacity, exclusive resources and verified cleanup.

Policy clears · model may holdView case → ExitCanary SaaS portability test interfaceSaaS portabilityExitCanarySubmitted

A synthetic CRM canary reveals what survives a vendor export before a buyer commits to the platform.

Model maps · evaluator decidesView case → AccessCrash outcome showing an unreachable process pathHuman reachabilityAccessCrashSubmitted

A capability-twin crash test finds where an apparently eligible person can no longer complete a real process.

Model compiles · code testsView case →
Open the complete Build Week collection
How to evaluate it

Ask what survived, not how many screens exist.

A useful portfolio should help you assess product judgment, not just visual taste.

  1. 01 · UserWho completes the real workflow?

    Look for a concrete user action, not a feature count.

  2. 02 · RiskWhat could block release?

    Data, roles, review, support and store constraints matter.

  3. 03 · EvidenceWhat proves the flow works?

    Device behavior, tests, release state and operating receipts.

  4. 04 · ControlWhat can another owner inherit?

    Code, accounts, known limits and handoff stay understandable.

Use the closest proof

Your product does not need to look like these.

Name the closest challenge and what is different. The first reply will say what carries over, what does not and which small step is enough.

Send five factstiger@ev1labs.com