What did Hlido score Clark Labs?

Clark Labs scored 66/100 (FADING) in Hlido's independent, hands-on review.

Does any vendor pay Hlido for placement?

No. Hlido takes no money from the agents it rates — scoring weights stay private and the evidence behind every verdict is public.

Clark Labs

Name: Clark Labs review
Item: Clark Labs
Rating: 66
Author: Hlido Editor

AI Agent · tested 2026-08-23 · re-test due 2026-11-23 · by the Hlido desk, not the vendor

In short: Two credible products — a cloud agent that keeps working with the tab closed and a memory-carrying desktop coding agent — wrapped in a lab claim nobody outside can check.

4 PASS · 1 FAIL of 5 public-surface claims

Visit Clark Labs → scorecard.json

Quick answer

Clark Labs scores 66/100 (FADING) on Hlido’s independent, hands-on test (reviewed 2026-08-23). FADING (66) because the two products are concrete and the release engineering is genuinely good — signed and notarized macOS, signed Windows, Linux packages — but the differentiating claim about the company (its own work Pricing was not findable on the public surface when tested.

Hlido Editor · 2026-08-23

Clark Labs ships a pair. Clark Agent runs on Clark's own cloud computer with a browser, terminal and file storage, taking a job and working it for hours while you are offline: research, site and document building, form filling, recurring workflows, with the result presented for review. Clark Code runs on your machine or a remote host over SSH, reads and edits your real project, holds persistent memory of the codebase's architecture and conventions, and can fan out to cloud research when it gets stuck. Both are concrete propositions, and the distribution is handled properly — macOS signed and notarized, Windows signed and released independently, Linux with .deb and .rpm packages. That is more distribution discipline than most agent startups manage, and it is the direct contrast with several peers in this batch. Two things hold the score down. The first is the framing: Clark describes itself as "the AI lab whose own engineering and research run as autonomous AI loops", which is a strong differentiating claim about how the company operates and is unverifiable from outside by construction. The second is commercial opacity — the surface says Clark Code is "priced so you can leave it running all day" without publishing a number, and the reviewed capture surfaced no pricing detail. "Cheap enough" is not a price. iOS is listed as coming soon and the Android/web split is stated honestly. This is a young lab with real engineering hygiene and a marketing register that outruns what a buyer can confirm.

Why FADING

FADING (66) because the two products are concrete and the release engineering is genuinely good — signed and notarized macOS, signed Windows, Linux packages — but the differentiating claim about the company (its own work running as autonomous AI loops) is unfalsifiable from outside, and pricing is described qualitatively rather than published. Comfortably clear of FLATLINE. It moves into STEADY when a number replaces 'priced so you can leave it running all day' and a third party can corroborate the operating claim.

Public-surface checklist

PASS Homepage loads (required)
PASS Primary value prop (required) — An AI agent in the cloud. A coding agent on your machine.
PASS Cta present (required) — Launch Clark Agent
FAIL Pricing or access — No price published; access described only as 'Priced so you can leave it running all day'
PASS Evidence or demo — 1 screenshot captured; in-page product walkthroughs for both Clark Agent and Clark Code

What we saw

1 screenshot captured by the Hlido engine during the reviewed run (run-7445912a1e5bf0c2-clarkchat-com). Our own captures — not vendor marketing material.

Clark Labs — run screenshot 1 (home.png) — home.png

What it does well

Two clearly separated products with distinct, concrete jobs rather than one blurred offering
Clark Agent continues working with the tab closed — a real differentiator, plainly stated
Clark Code holds persistent memory of repo architecture and past decisions
macOS build is signed and notarized; Windows signed; Linux ships .deb and .rpm
Runs on your machine or over SSH — no forced cloud sandbox for the coding agent
Platform availability stated honestly, including iOS as not yet shipped

What it fails at

No published price for either product — 'priced so you can leave it running all day' is not a number
The 'our own engineering runs as autonomous AI loops' claim cannot be verified from outside
No named customers, case studies, or usage evidence on the surface
No API, MCP server or SDK surfaced for programmatic integration
Young lab; no continuity or funding signal published
iOS unavailable, limiting the mobile story the cloud agent implies

Red flags

Central differentiating claim — that the lab's own engineering and research run as autonomous AI loops — is unverifiable from any public surface

Best for

Individuals wanting a cloud agent that finishes long-running errands unattended
Developers who want a coding agent with persistent codebase memory running locally or over SSH
Users on Linux or Windows underserved by macOS-only agent tooling

Not recommended for

Buyers who need a published price before evaluating
Teams requiring programmatic integration — there is no documented API or MCP surface
Procurement processes needing customer references or a continuity signal
iOS-first users

Pricing & access

Pricing findable on the public surfaceFAIL No price published; access described only as 'Priced so you can leave it running all day' (tested 2026-08-23)

Derived from Hlido-held evidence only (engine checklist + editorial text); quotes are verbatim from the scorecard; not vendor-supplied; re-derived daily. Verify current prices on the vendor's pricing page. Last verified 2026-08-23.

Compared to

Cursor persistent-repo-memory
Cursor is the established editor-based coding agent with a published price and large user base; Clark Code adds persistent repo memory and an SSH path but publishes no pricing. Cursor for a known quantity, Clark for the memory model and remote-host support.
Aider packaged-vs-open
Aider is a mature open CLI agent you can inspect end to end; Clark Code is a closed native app with a research-fanout backend. Aider for transparency and zero lock-in, Clark for the packaged desktop experience.

Agent relevance

No programmatic surfaces

Agentic-Commerce Readiness 30/100 · SURFACE-ONLY

Independent readiness for agent delegation & transaction. How it’s scored · check live

Both products are end-user applications — a hosted web/Android agent and a native desktop coding agent. No API, MCP server, SDK or webhook is documented on the public surface, so an agent cannot discover or drive Clark programmatically. Clark Code itself consumes tools, but does not expose any.

Agent-friendly score: 2/10

Evidence

Clark Agent runs on Clark's cloud computer with browser, terminal and file storage — source (2026-08-23) verified
Continues working while the user is offline with the tab closed — source (2026-08-23) verified
Clark Code runs locally or over SSH with persistent codebase memory — source (2026-08-23) verified
macOS build is signed and notarized; Windows signed; Linux .deb and .rpm available — source (2026-08-23) verified
No published price for Clark Code — source (2026-08-23)
Claims the lab's own engineering and research run as autonomous AI loops — source (2026-08-23)

scorecard.json · transparency passport · registry · methodology

More: compare agents · best of · developer tools · incident registry

Verdict by Hlido Editor, our automated editorial system · Method: public-surface-tier-1+editorial-narrative-v2 · Methodology version 2026.05 · Next review due 2026-11-23

How this page was produced. The scores, claim verdicts and evidence come from automated hands-on testing of the product’s public surface. The written analysis is drafted by an AI system, and pages publish without a person reviewing each one. Hlido publishes this record and answers for it — tell us if anything here is wrong and we will correct it.

Embed this trust badge

Live, always-current independent score — free to embed on your site or README. No vendor pays for placement.

Markdown

[![Hlido trust score](https://hlido.eu/badge/clarkchat.svg)](https://hlido.eu/check/?agent=clarkchat)

HTML

<a href="https://hlido.eu/check/?agent=clarkchat"><img src="https://hlido.eu/badge/clarkchat.svg" alt="Hlido trust score"></a>