Skip to content
View Grit-77's full-sized avatar
💭
💭
2 followers · 3 following

Achievements

Achievement: Pair Extraordinairex2Achievement: Pull Sharkx3Achievement: QuickdrawAchievement: YOLO

Achievements

Achievement: Pair Extraordinairex2Achievement: Pull Sharkx3Achievement: QuickdrawAchievement: YOLO

Highlights

  • Pro

Block or report Grit-77

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Grit-77/README.md

Grit. Proof stands in the light. AI systems that check their own work. A limestone hall with the Grit disk drawn on the floor as an engineering overlay: three rings of tiles around a hub, labelled like a technical figure.

Website · RADAR · How RADAR works · Company · Contact

Grit builds AI systems that check their own work. An AI agent that says done has made a claim; everything we build closes on a check that runs again, not on the claim. Our product is RADAR. For companies and individuals we also build websites, AI automations and autonomous systems, and bring scattered data into one place.

RADAR

RADAR. A claim is not a result; RADAR reruns the test. Runs Claude Code, Codex CLI, OMP and Hermes Agent. A radio dish at blue hour inside radar-scope rings, with a sample readout: verify pass, attempt 1 of 3, aground 0.

RADAR runs the AI coding agents you already use (Claude Code, Codex CLI, OMP and Hermes Agent) and closes a task only after it reruns the acceptance test itself. It grew out of the infrastructure we use for our own work.

The test is written first → an agent works in its own copy of the project → it says done, recorded as a claim → RADAR reruns the test → it passes: the task closes and the proof is kept · it fails: the task reopens with the failure attached.

RADAR's chart room, its home screen in a terminal, with sample data: a chart of machines and the busy worker slots on each, the LOG of tasks, NEXT for what needs a person, ARRIVALS per hour, token use per machine and the COMMANDS list.

The chart room, RADAR's home screen in a terminal, with sample data.

In Grit's own work (20-28 Sep 2026), workers running one coding agent reported a task as finished 1,608 times (1,457 tasks); Grit's immediate re-run of the acceptance commands on the same server failed on 74 of those reports (4.6%). A claim alone does not close a task.

Grit internal ledger, snapshot 28 Sep 2026 06:00 UTC. Observed, not a controlled comparison.

The website draws the figure, one square per report.

  • For individuals: an open edition, open source (Apache-2.0) at public release. In early access now: ask for a place.
  • For companies: a company edition, in development. Talk to us.

What we do

  • Websites, designed and built with you from the first draft to the live site: accessible, quick to load, in your visitors' languages.
  • AI automations: repetitive work done the same way each time, every result checked before it is used.
  • Autonomous systems: agents that carry a piece of work from start to finish; a task closes only when its test passes again.
  • Data: scattered data brought into one place and cleaned, reliable enough to decide on.

Write to us

radarbygrit@gmail.com · X @gritradar_ai · Instagram @gritradar.ai

Grit · Ankara · founded June 2026 · grit.grit-77.workers.dev

Popular repositories Loading

  1. Grit-77 Grit-77 Public

    Grit's public profile: who we are, what we build, and RADAR, the product that closes a task only when its test passes again.