{
  "schema_version": "2.0",
  "slug": "construct-computer",
  "name": "Construct",
  "agent_url": "https://construct.computer",
  "category": "Workflow & Automation",
  "run_id": "run-r4-v2-construct-computer-2026-08-27",
  "run_at": "2026-08-27T04:31:27.122Z",
  "editor": "Hlido Editor",
  "editorial_method": "public-surface-tier-1+editorial-narrative-v2",
  "methodology_version": "2026.05",
  "methodology_url": "/methodology/public-surface-tier-1/",
  "score": 76,
  "tier": "STEADY",
  "laddoo_score": 76,
  "confidence": "medium-high",
  "hlido_opinion": {
    "headline": "An \"AI employee\" for solo founders and small teams that bundles workflows, scheduled jobs, a cloud desktop and inspectable memory — with fully published pricing from $9/month.",
    "body": "Construct is aiming at the part of the agent market that actually converts: a solo founder or a small team that wants recurring work done, not a framework to build with. The product is a bundle rather than a single capability — encode a process once as agent steps with connected apps and notifications, version it, schedule it, and let anyone on the team re-run it; describe an internal tool and have Construct write, validate and publish a working app into your cloud desktop; schedule agent prompts or whole workflows and keep the result history; and reach it from web, Slack, Telegram, Discord slash commands or its own email inbox. The feature we would single out is memory you can inspect: preferences, decisions and project context are stated to be kept with evidence and history, and explicitly reviewable, correctable and forgettable. Most agent products treat memory as an opaque quality improvement; making it auditable is the harder and more honest choice, and it is the right one for a tool that accumulates knowledge about your business. Pricing is fully published across Lite ($9/mo, $7.50 annual), Starter ($59/mo, $39 annual) and higher tiers, with concrete limits stated per tier — agent count, steps per task, parallel background jobs, storage, command runtime — rather than the usual unquantified 'usage'. That is unusually specific and lets a buyer size the plan before signing up. Against that: every capability claim here is vendor-described marketing, the review captured no measured run, and 'AI employee' is a framing that promises more autonomy than any public surface can evidence. Routing and access policies are noted to vary by channel without those differences being spelled out.",
    "voice": "Hlido Editor",
    "as_of": "2026-08-27",
    "editor_signature_pending": true
  },
  "tier_rationale": "STEADY (76) because pricing is fully published with concrete per-tier limits (agents, steps, parallel jobs, storage, runtime) rather than vague usage bands, there is a real free trial, and inspectable, correctable memory is a genuine transparency choice rather than a feature bullet. Held mid-band because the entire capability set is vendor-described with no measured run in evidence, the \"AI employee\" framing overstates what any public surface can support, and per-channel routing and access differences are flagged as varying without being documented.",
  "what_it_does_well": [
    "Fully published pricing with concrete per-tier limits — agent count, steps per task, parallel jobs, storage and command runtime",
    "Memory is explicitly inspectable, correctable and forgettable, with evidence and history retained",
    "Processes are encoded once, versioned, scheduled and re-runnable by anyone on the team",
    "Reachable from web, Slack, Telegram, Discord slash commands and a native email inbox",
    "Builds and publishes working internal tools into a cloud desktop rather than only returning text",
    "Workspace model with invitations, roles and access controls rather than single-user only",
    "7-day Pro trial and plans from $9/month lower the trial cost to near zero"
  ],
  "what_it_fails_at": [
    "Every capability claim is vendor-described; no measured run is in evidence on this review",
    "\"AI employee\" promises a degree of autonomy no public surface can substantiate",
    "Per-channel routing and access policies are stated to vary without those differences being documented",
    "Research output is described as cited but the citation quality is not demonstrated",
    "Lower tiers are tightly bounded (2 agents, 50 steps per task, 5-minute runtime on Lite) — the entry price buys a small envelope"
  ],
  "best_for": [
    "Solo founders and small teams that want recurring processes run, not a framework to build on",
    "Teams that need agent memory they can audit and correct rather than trust blindly",
    "Workflows already living in Slack, Telegram, Discord or email",
    "Buyers who want to size a plan against published, concrete limits before committing"
  ],
  "not_recommended_for": [
    "Engineering teams that want a programmable framework and their own orchestration",
    "Workloads exceeding the published per-tier step, runtime or parallelism limits without moving up several tiers",
    "Buyers who need independently verified autonomy before delegating real work"
  ],
  "red_flags": [],
  "compared_to": [
    {
      "slug": "lindy",
      "verdict_diff": "Both sell an AI-assistant-as-teammate to small teams. Construct adds a cloud desktop, internal-tool generation and explicitly inspectable memory, and publishes concrete per-tier execution limits. Lindy is the more established assistant-and-workflow product. Construct if auditable memory and published limits matter; Lindy for maturity.",
      "preferred_for_axis": "inspectable-memory-and-published-limits-vs-maturity"
    },
    {
      "slug": "n8n",
      "verdict_diff": "n8n is a developer-facing workflow engine you assemble yourself; Construct encodes processes conversationally and runs them as an assistant for non-engineers. n8n for control and self-hosting, Construct for teams with no one to build the pipeline.",
      "preferred_for_axis": "assistant-for-non-engineers-vs-developer-workflow-engine"
    }
  ],
  "evidence_urls": [
    {
      "claim": "Positioned as an AI employee for solo founders and small teams; plans from $9/month with a 7-day Pro trial",
      "source": "https://construct.computer ('An AI employee you can actually work with'; '7 days of Pro free. Then plans from $9/month. Cancel any time.')",
      "tested_at": "2026-08-27",
      "verified": true
    },
    {
      "claim": "Published tiers with concrete limits: Lite $9/mo (2 agents, 50 steps per task, 2 parallel jobs, 100 MB storage, 5-minute runtime); Starter $59/mo",
      "source": "https://construct.computer ('Lite $9 / month … Up to 2 AI agents / Up to 50 steps per task / 2 parallel background jobs / 100 MB cloud storage / 5-minute command runtime'; 'Starter $59')",
      "tested_at": "2026-08-27",
      "verified": true
    },
    {
      "claim": "Memory is inspectable, correctable and forgettable with evidence and history",
      "source": "https://construct.computer ('Preferences, decisions, and project context stay with evidence and history. Review, correct, or forget anything it learns so nothing important lives only in your head.')",
      "tested_at": "2026-08-27",
      "verified": true
    },
    {
      "claim": "Reachable from web, Slack, Telegram, Discord slash commands and a native email inbox, with routing varying by channel",
      "source": "https://construct.computer ('Message Construct from the web, Slack, Telegram, Discord slash commands, or its native email inbox. Routing and access policies vary by channel.')",
      "tested_at": "2026-08-27",
      "verified": true
    }
  ],
  "agent_relevance": {
    "has_api": false,
    "has_cli": false,
    "has_mcp": false,
    "has_webhook": false,
    "has_sdk": false,
    "behavioral_testable": false,
    "agent_integration_path": "Construct is a destination product rather than a component. It is driven by humans through web, Slack, Telegram, Discord or email, and connects outward to apps you authorise. No public API, MCP server, SDK or webhook surface is advertised on the reviewed page, so another agent cannot drive it programmatically — you delegate to it, you do not build on it.",
    "agent_friendly_score": 3
  },
  "marking_signal": {
    "marking_statement": null,
    "detection_tool": null,
    "cop_signatory": null,
    "evidence_url": null,
    "checked_at": "2026-08-27",
    "note": "Construct drafts emails, reports and research that reach third parties, so output marking is a live question for it — but the reviewed public surface makes no marking, provenance or disclosure claim. Recorded as unevidenced, not as absent."
  },
  "checklist": [
    {
      "id": "homepage_loads",
      "pass": true,
      "required": true,
      "tested_at": "2026-08-27"
    },
    {
      "id": "primary_value_prop",
      "pass": true,
      "required": true,
      "evidence": "'An AI employee you can actually work with'",
      "tested_at": "2026-08-27"
    },
    {
      "id": "cta_present",
      "pass": true,
      "required": true,
      "evidence": "'Start Now' / 'Book a walkthrough'",
      "tested_at": "2026-08-27"
    },
    {
      "id": "pricing_or_access",
      "pass": true,
      "required": true,
      "evidence": "Full tier table with concrete limits: Lite $9/mo, Starter $59/mo, annual discounts, 7-day Pro trial",
      "tested_at": "2026-08-27"
    },
    {
      "id": "memory_transparency",
      "pass": true,
      "required": true,
      "evidence": "'Review, correct, or forget anything it learns'",
      "tested_at": "2026-08-27"
    }
  ],
  "evidence_images": {
    "run_id": "run-22618422dc965807-construct-computer",
    "base": "https://images.hlido.eu/reviews/construct-computer/run-22618422dc965807-construct-computer",
    "files": [
      "home.png",
      "page__main.png",
      "page_.png",
      "page_pricing_.png"
    ],
    "urls": [
      "https://images.hlido.eu/reviews/construct-computer/run-22618422dc965807-construct-computer/home.png",
      "https://images.hlido.eu/reviews/construct-computer/run-22618422dc965807-construct-computer/page__main.png",
      "https://images.hlido.eu/reviews/construct-computer/run-22618422dc965807-construct-computer/page_.png",
      "https://images.hlido.eu/reviews/construct-computer/run-22618422dc965807-construct-computer/page_pricing_.png"
    ]
  },
  "pricing_facts": {
    "schema": "pricing-facts/1",
    "model": [
      "subscription"
    ],
    "price_points": [
      {
        "quote": "…a cloud desktop and inspectable memory — with fully published pricing from $9/month.",
        "source": "hlido_opinion.headline"
      },
      {
        "quote": "…knowledge about your business. Pricing is fully published across Lite ($9/mo, $7.50 annual), Starter ($59/mo, $39 annual) and higher tiers, with concrete limits…",
        "source": "hlido_opinion.body"
      }
    ],
    "pricing_disclosed": {
      "pass": true,
      "evidence": "Full tier table with concrete limits: Lite $9/mo, Starter $59/mo, annual discounts, 7-day Pro trial",
      "tested_at": "2026-08-27"
    },
    "last_verified": "2026-08-27",
    "basis": "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.",
    "derived_at": "2026-08-27"
  }
}