Does Laser send screenshots to the Jev API?

No. Images are OCR'd locally with grim and tesseract and are never saved or uploaded. What Jev sees depends on the privacy level: a site name, a bag of dictionary keywords, or masked screen text. `laser audit` shows each payload that was sent.

What happens if Jev misjudges a page as a distraction?

Clicking the nudge or right-clicking the reticle marks that site on-task for the rest of the session. Closing only happens after four minutes of continuous drift, only affects browser tabs or web apps, and Warn-only mode never closes anything.

What do I need beyond Omarchy to run Laser?

python3 3.11 or higher and hyprctl, both of which ship with Omarchy. The Balanced and Full privacy levels also need the grim and tesseract packages plus tesseract-data-eng for on-device OCR; without them Laser still works from site names.

When does Laser actually send a check to Jev?

Only during a session, and only when the window changes or a verdict goes stale, which it treats as 20 seconds when unsure and 60 seconds when clear. Nothing runs while idle, on a break, or while the screen is locked.

How much does a day of Laser cost?

Jev charges $0.042 per million input tokens and output is free, and a check is about 450 tokens of fixed question schema plus your data. The README estimates roughly $0.30 to $0.50 a month of daily use, with live cost shown in the panel.

How do I remove Laser cleanly?

Stop any running session first so your window border returns to your theme, then run omarchy plugin remove yasserdo.laser and delete ~/.config/laser, ~/.local/share/laser and ~/.local/state/laser. If a laser daemon survives, end it with pkill -f "laser daemon".

Does the escalation ever close native apps?

No. Only browser tabs and web apps are ever closed, and only after four minutes of continuous drift. Strict halves the timings, Warn only never closes anything, and Keep it hazy swaps closing for the window haze.

CommunityJev Integrations & Adapters

Laser

Focus mode for Omarchy and Hyprland that judges each screen against the task you declared and escalates from a red reticle to a window haze and, as a last resort, closing the drifting browser tab. Each check is one Jev on-task classification, with screenshots OCR'd locally and never uploaded.

See highlights

Open github.com

Stars: 6
Language: Python
License: MIT
Published by: Community
Last verified: Oct 2026

iYassr/omarchy-laser

User repository on GitHub

View on GitHub

Laser: AI focus mode & distraction blocker for Omarchy / Hyprland. Judges every screen against your task (TypeSafe Jev), privacy-first, ADHD-friendly, ~1¢/day.

Forks: 0
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.Homepage

About this project

Tell it what you're working on. It notices when you drift and pulls you back. An AI distraction blocker for Omarchy and Hyprland on Linux. It judges each screen against your task, not a blocklist, it's private by design, and it costs about a cent a day. You sit down to work and twenty minutes later you're in a WhatsApp thread or deep in Reddit, and you can't say how you got there. Classic site blockers don't fit how people actually work:

Quoted from the project's README. Read the full README.

Tagscommunity focus linux privacy

Works forJev for Classification Jev for Decision gating

Used inJev for Software & DevTools

Languages: Python
Platforms: CLI
Capabilities: Open sourceRuns locallyRequires API key

Highlights

Python standard library only; checks take about half a second and run through TypeSafe's Jev API.
Three privacy levels; screenshots are OCR'd on-device with grim and tesseract and never leave the machine.
Gentle escalation: red reticle at 30 s, haze over the window at 3 min, tab closed at 4 min; only browser tabs are ever closed.
Reported cost of roughly $0.30-0.50 per month of daily use; `laser audit` shows the exact JSON sent to Jev.
Tested on 14 real pages: no privacy level flagged real work as a distraction (Strict 11/14, Balanced 12/14, Full 14/14).

Watch out

MIT-licensed, but it needs Omarchy on Hyprland plus a paid TypeSafe API key, and the Balanced/Full privacy levels also need grim and tesseract for on-device OCR.

Questions about this entry

Short answers drawn from the linked project or page.

Related terms

Glossary definitions related to this entry.

Zero data retention

More Jev resources like this

Qualm

macOS menu bar screen-time app that reads the screen as text through the Accessibility API and asks a System One model one Choice per user rule plus a Noul on whether the page is a payment, login or banking screen, stepping in only when a rule's probability clears its threshold.

Jev Integrations & Adapters#community#macos#digital-wellbeing

Community

RoderickQiu/qualmPython

Pydantic AI

20.5k

Python agent framework whose TypeSafeModel class runs Jev as a decision model inside Pydantic AI agents: Pydantic output-type fields become typed Jev questions answered together in one request, each answer carrying its own confidence, and tools can be offered as Choice options.

Jev Integrations & Adapters#community#agents#framework

Community

pydantic/pydantic-aiPython

IntentSQL

Alpha playground that answers plain-English questions over SQLite by composing small Jev decisions -- table, column, comparison and operand selection -- into inspectable typed plans, deterministic SQL and previewed guarded writes. Bundles Cyberchase, DESE and Moneyball databases.

Jev Integrations & Adapters#community#sqlite#text-to-sql

Community

Amine-LG/IntentSQLPython

SignalChain

Data-cleaning framework that asks Jev Choice questions to classify each CSV column into a 13-code type vocabulary and each dataset into one of six scenes, then executes every write locally; an optional System 2 LLM handles the default path or escalations.

Jev Integrations & Adapters#community#data-cleaning#pandas

Community

Cogniform/AI-decision-makerPython

Jevidence

Python sandbox and tutorial companion that turns Jev Choice and Noul judgments about issue category and reproduction into code-applied queue proposals, with confidence gates, a review fallback, synthetic offline fixtures and labeled replay. Live mode is opt-in.

Jev Guides & Articles#community#issue-triage#guardrails

Community

peakevergreen/jevidencePython

AutoGPT

187.5k

Autonomous agent platform with 187,504 stars for building and running agents from plain English or a visual builder. Its platform features first-class TypeSafe Jev decision blocks for typed routing, filtering, scoring and confidence-gated next-action dispatching.

Jev Integrations & Adapters#community#autonomous-agents#platform

Community

Significant-Gravitas/AutoGPTPython

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai