{
  "data": {
    "a": {
      "slug": "aura-autopilot",
      "name": "Aura Autopilot",
      "vendor": "Ramen",
      "vendorUrl": "https://www.tryaura.dev/autopilot/",
      "company": "tryaura.dev",
      "kind": "mcp",
      "category": "desktop-automation",
      "summary": "Aura Autopilot is Ramen's Apache-2.0 game playtester for Unreal Engine 5, Unity 6 and Godot 4. An editor plugin and local agent service play the game, and Claude Code or Codex drive it through a stdio MCP server.",
      "url": "https://www.anchorterminal.com/tools/aura-autopilot",
      "markdownUrl": "https://www.anchorterminal.com/tools/aura-autopilot.md",
      "slimMarkdownUrl": "https://www.anchorterminal.com/tools/aura-autopilot.min.md",
      "jsonUrl": "https://www.anchorterminal.com/api/v1/tools/aura-autopilot.json",
      "dns": "aura-autopilot.grade.anchorterminal.com",
      "repo": "https://github.com/ramenvr/autopilot",
      "license": "Apache-2.0",
      "transports": [
        "stdio"
      ],
      "packages": [],
      "auth": "api-key",
      "authNotes": "No Ramen account needed with one's own keys. A run needs a TypeSafe API key for the Jev decision model and a model provider key or Claude or ChatGPT plan, or an Aura sign-in in the browser instead. Keys are kept in autopilot.env in the user settings folder.",
      "pricing": "byo-plan",
      "pricingNotes": "Free under Apache-2.0. Runs pay for Jev through a TypeSafe key ($0.042 per million input tokens, per our typesafe-jev listing) or an Aura plan from $10 a month billed annually, plus the coding agent's model on the user's key or plan.",
      "priceSummary": "Your plan",
      "where": "local",
      "x402": {
        "level": "no",
        "evidence": "No x402, MPP or L402 in the repository, the docs or the download page, checked 1 November 2026.",
        "endpoints": []
      },
      "toolCount": 14,
      "popularity": {
        "githubStars": null,
        "npmWeekly": null,
        "pypiWeekly": null,
        "asOf": "2026-11-01"
      },
      "docsUrl": "https://www.tryaura.dev/documentation/autopilot",
      "capabilities": [
        "desktop.control",
        "desktop.scripting"
      ],
      "tags": [
        "open-source",
        "apache-2.0",
        "local",
        "mcp",
        "stdio",
        "python",
        "game-testing",
        "unreal-engine",
        "unity",
        "godot",
        "claude-code-plugin",
        "pre-1.0"
      ],
      "graded": true,
      "anchor": {
        "graded": true,
        "score": 58.7,
        "grade": "C",
        "agentReady": false,
        "rank": 840,
        "ranked": true,
        "rankOf": 1557,
        "categoryRank": 4,
        "methodology": "1.0",
        "run": "2026-10-01",
        "scores": {
          "ergonomics": 69,
          "maintenance": 43,
          "payments": 75,
          "reliability": 41,
          "schema": 74,
          "security": 53,
          "transparency": 55
        },
        "assessment": {
          "confidence": "medium",
          "date": "2026-11-01"
        },
        "negative": 0,
        "verdict": "Apache-2.0, with typed MCP tools carrying read-only hints, a separate process for model-written code and headless exit codes for CI. It is days old with no changelog, needs a TypeSafe Jev key or an Aura account, and by default shares recordings, agent traces and the user's own agent conversation with Ramen unless turned off.",
        "bestFor": "Agent-driven playtests of Unreal, Unity or Godot games in the editor or in CI",
        "strengths": [
          "Apache-2.0 source for the agent service, the Unreal, Unity and Godot plugins and the Claude Code plugin in one repository",
          "Fourteen MCP tools with JSON Schema inputs, readOnlyHint on the seven that only read, and a `wait` call for long episodes",
          "Model-written harness and probe code runs in a child process without API keys; on Linux bubblewrap also removes network and most disk access",
          "Headless runs from the engine's command line write result.json and exit 0, 1, 2 or 3 for CI",
          "Release downloads are signed, listed in SHA256SUMS.txt and published as immutable GitHub releases"
        ],
        "weaknesses": [
          "Run sharing is on by default and sends recordings, screenshots, agent traces with code excerpts and, for MCP runs, the user's Claude Code or Codex conversation to app.tryaura.dev",
          "Shared run data already sent is never deleted, even by `autopilot uninstall`, and no retention period was found",
          "Python package 0.1.0 and an Unreal plugin flagged beta, with no changelog in the repository; docs/run-data.md and docs/architecture.md are referenced but not published",
          "Needs a TypeSafe key for the Jev decision model (early access, per our typesafe-jev listing) or an Aura account, plus a model provider key or plan",
          "On Windows and macOS, the supported platforms, the sandbox is a plain child process that keeps keys out but not network or disk access"
        ],
        "agentNotes": [
          "Check the game's project is open in the editor with the Autopilot tab connected before calling start_run; otherwise it answers that the agent service isn't running.",
          "When a call answers \"still going\", call `wait` until it returns; don't start another tool call meanwhile.",
          "Pass the user's request in their own words with kind verify, goal or explore, and read the INSTRUCTIONS.md that start_run names in parts with Autopilot's read_file.",
          "Tell the user that run data is shared by default and how to turn it off (the Settings page, AUTOPILOT_SHARE=0 or -PlaytestShare=0) before the first run.",
          "Never ask the user to paste a TypeSafe or provider key into the conversation; keys belong in autopilot.env or the Autopilot tab."
        ],
        "metrics": {
          "kind": "local",
          "measured": false
        },
        "reviewCount": 0,
        "avgRating": 0,
        "history": [
          {
            "basis": "public evidence",
            "confidence": "medium",
            "grade": "C",
            "methodology": "1.0",
            "run": "2026-10-01",
            "runLabel": "October 2026 research run",
            "score": 58.7
          }
        ],
        "editorialScores": {
          "ergonomics": 69,
          "maintenance": 43,
          "payments": 75,
          "reliability": 41,
          "schema": 74,
          "security": 53,
          "transparency": 58
        },
        "provenanceScore": 52,
        "checked": {
          "categories": {
            "ergonomics": "2026-11-01",
            "maintenance": "2026-11-01",
            "payments": "2026-11-01",
            "reliability": "2026-11-01",
            "schema": "2026-11-01",
            "security": "2026-11-01",
            "transparency": "2026-11-01"
          },
          "last": "2026-11-01",
          "lastMoved": "2026-11-13",
          "oldest": "2026-11-01",
          "stale": false,
          "staleAfterDays": 90
        }
      },
      "connect": {
        "install": "# Download the plugin for your engine and OS from https://www.tryaura.dev/autopilot/ and finish the first-run setup in the editor's Autopilot tab\nautopilot connect claude\nautopilot connect codex",
        "claudeCode": "claude plugin install autopilot --marketplace ramenvr/autopilot",
        "config": {
          "mcpServers": {
            "autopilot": {
              "args": [
                "mcp"
              ],
              "command": "autopilot"
            }
          }
        }
      },
      "letme": {
        "capability": "https://letme.dev/desktop.control",
        "tool": "https://letme.dev/aura-autopilot"
      },
      "area": "developer",
      "unitPrices": [
        {
          "item": "Aura Indie, billed monthly",
          "unit": "month",
          "usd": 20,
          "note": "Includes Autopilot; $10 a month billed annually ($120 a year). Two-week trial, no card"
        },
        {
          "item": "Aura Pro",
          "unit": "month",
          "usd": 40,
          "note": "Includes Autopilot; $360 a year billed annually"
        },
        {
          "item": "Aura Ultimate",
          "unit": "month",
          "usd": 200,
          "note": "Includes Autopilot; $1,800 a year billed annually"
        }
      ],
      "provenance": {
        "legalEntity": "TribeVR Inc. (named in Aura's terms of service and privacy policy), trading as Ramen",
        "domain": "tryaura.dev",
        "domainRegistered": "2025-04-29",
        "endpointOnVendorDomain": null,
        "terms": "https://www.tryaura.dev/terms-of-service",
        "privacy": "https://www.tryaura.dev/privacy-policy",
        "statusPage": "",
        "changelog": "",
        "securityTxt": "none",
        "checked": "2026-11-01",
        "notes": [
          "The repository's licence file is Apache 2.0; Aura's terms (25 June 2026) and privacy policy (27 August 2026) are TribeVR Inc.'s and don't mention Autopilot.",
          "RDAP gives tryaura.dev a registration date of 29 April 2025 and ramenvr.com of 16 February 2019.",
          "www.tryaura.dev/.well-known/security.txt answered 404. SECURITY.md in the repository asks for GitHub private vulnerability reporting.",
          "app.tryaura.dev, where shared run data goes and where sign-in happens, disallows ClaudeBot, anthropic-ai and Claude-Web in robots.txt and carries Content-Signal ai-train=no, so nothing on it was read.",
          "Local software, so there is no status page. Aura's release notes mention Autopilot; the repository has no changelog."
        ],
        "score": 52
      },
      "pageJsonUrl": "https://www.anchorterminal.com/tools/aura-autopilot.json",
      "live": {
        "slug": "aura-autopilot",
        "versions": [
          {
            "registry": "github",
            "name": "ramenvr/autopilot",
            "version": "v1.0.0-1ebf2a5",
            "released": "2026-11-16",
            "seenAt": "2026-11-18T17:29:36.158987378Z"
          }
        ],
        "githubStars": 875,
        "securityTxt": {
          "url": "https://tryaura.dev/.well-known/security.txt",
          "state": "none",
          "checkedAt": "2026-11-18T16:29:23.39474166Z"
        },
        "domain": {
          "domain": "tryaura.dev",
          "registered": "2025-04-29",
          "source": "https://pubapi.registry.google/rdap/domain/tryaura.dev",
          "checkedAt": "2026-11-15T15:36:25.308323945Z"
        },
        "pages": [
          {
            "url": "https://www.tryaura.dev/privacy-policy",
            "kind": "privacy",
            "status": 304,
            "checkedAt": "2026-11-18T22:10:48.459775279Z",
            "changedAt": "0001-01-01T00:00:00Z",
            "fingerprint": "69fdb3dbd602"
          },
          {
            "url": "https://www.tryaura.dev/terms-of-service",
            "kind": "terms",
            "status": 304,
            "checkedAt": "2026-11-18T22:10:50.526594628Z",
            "changedAt": "0001-01-01T00:00:00Z",
            "fingerprint": "9edeed555c94"
          }
        ],
        "github": {
          "repo": "ramenvr/autopilot",
          "checkedAt": "2026-11-16T11:35:11.742545749Z",
          "stars": 866,
          "archived": false,
          "issuesEnabled": true,
          "pushedAt": "2026-11-16T05:03:22Z",
          "openIssues": 3,
          "openPRs": 5,
          "issuesOpened90d": 6,
          "issuesClosed90d": 3,
          "prsMerged90d": 63,
          "ci": {
            "branch": "main",
            "runs": 26,
            "success": 22,
            "failure": 4,
            "other": 0,
            "latest": "success",
            "latestAt": "2026-11-13T16:54:03Z"
          },
          "replies": {
            "sampled": 6,
            "unanswered": 0,
            "medianDays": 1.9
          }
        },
        "updatedAt": "2026-11-18T22:10:50.526594628Z"
      }
    },
    "answer": "Aura Autopilot scores 58.7 (C) on agent readiness against Glasshand's 50.8 (D), and leads in 5 of 7 scored categories. Glasshand leads on schema \u0026 documentation and security \u0026 auth.",
    "b": {
      "slug": "glasshand",
      "name": "Glasshand",
      "vendor": "Codapt Inc.",
      "vendorUrl": "https://glasshand.com",
      "company": "glasshand.com",
      "kind": "http-api",
      "category": "desktop-automation",
      "summary": "Glasshand rents dedicated physical iPhones and Mac minis in a data centre by the month, and an agent controls them through an HTTP API that returns screenshots and sends taps, clicks and keystrokes. Run by Codapt Inc.",
      "url": "https://www.anchorterminal.com/tools/glasshand",
      "markdownUrl": "https://www.anchorterminal.com/tools/glasshand.md",
      "slimMarkdownUrl": "https://www.anchorterminal.com/tools/glasshand.min.md",
      "jsonUrl": "https://www.anchorterminal.com/api/v1/tools/glasshand.json",
      "dns": "glasshand.grade.anchorterminal.com",
      "license": "Proprietary hosted service under the Glasshand Terms of Service",
      "transports": [
        "http"
      ],
      "remoteUrl": "https://api.glasshand.com/v1",
      "packages": [],
      "auth": "oauth",
      "authNotes": "OAuth device authorisation. An agent calls `/auth/device/code` with no credential, and the person signs in by email code and approves the named agent and its scopes (`rentals:read`, `rentals:write`, `devices:control`, `billing:write`). Access tokens last 15 minutes and refresh tokens rotate. The person can revoke an agent from the dashboard's Agents page, which rejects its requests at once.",
      "pricing": "paid",
      "pricingNotes": "$60 a month for an iPhone SE (3rd generation) and $90 a month for a Mac mini M4 with 16 GB and 256 GB, plus tax, each rental its own monthly subscription through Stripe. No free tier, trial, hourly or daily rental. If setup fails within 30 minutes of payment the first payment is refunded in full. An eSIM plan for the iPhone is added separately, and its price wasn't found.",
      "priceSummary": "$60 / mo",
      "where": "hosted",
      "x402": {
        "level": "no",
        "evidence": "No x402, MPP or L402 in agents.md, the OpenAPI document, the pricing page or the terms (checked 2026-11-18). Checkout returns a Stripe link for the person to pay.",
        "endpoints": []
      },
      "toolCount": null,
      "popularity": {
        "githubStars": null,
        "npmWeekly": null,
        "pypiWeekly": null,
        "asOf": "2026-11-18"
      },
      "docsUrl": "https://glasshand.com/agents.md",
      "openapi": "https://api.glasshand.com/openapi.json",
      "capabilities": [
        "mobile.control",
        "desktop.control"
      ],
      "tags": [
        "hosted",
        "paid",
        "subscription",
        "openapi",
        "oauth",
        "ios",
        "macos",
        "idempotency"
      ],
      "lastRelease": "2026-10-22",
      "graded": true,
      "anchor": {
        "graded": true,
        "score": 50.8,
        "grade": "D",
        "agentReady": false,
        "rank": 1143,
        "ranked": true,
        "rankOf": 1557,
        "categoryRank": 7,
        "methodology": "1.0",
        "run": "2026-10-01",
        "scores": {
          "ergonomics": 62,
          "maintenance": 35,
          "payments": 35,
          "reliability": 27,
          "schema": 79,
          "security": 60,
          "transparency": 52
        },
        "assessment": {
          "confidence": "medium",
          "date": "2026-11-18"
        },
        "negative": 0,
        "verdict": "Glasshand gives an agent a real iPhone or Mac through a small OpenAPI-described API with OAuth device sign-in, four scopes, short-lived tokens and idempotency keys on every write and input. Rentals are monthly with no trial, and no status page, SLA, changelog or security programme was found.",
        "bestFor": "Agents that need a real iPhone or Mac, such as testing native apps in Xcode or using iOS-only apps and a real phone number.",
        "strengths": [
          "OAuth device flow that an agent starts itself, with four scopes, 15-minute access tokens and rotating refresh tokens",
          "An `Idempotency-Key` on checkout, cancel, resume and each input, with written retry rules",
          "A public OpenAPI 3.1 document and agent instructions in Markdown at agents.md",
          "Each device is dedicated to one customer and is wiped and factory reset on release",
          "Agent instructions state that screens, web pages and incoming messages are untrusted task data"
        ],
        "weaknesses": [
          "No status page, SLA, changelog or security.txt found",
          "Monthly rental only, from $60, with no free tier or trial",
          "Control is screenshots and input only, with no shell, file or clipboard API, and text input is limited to US ASCII",
          "`devices:control` grants full input to every device the agent can reach, with no per-action confirmation",
          "The home page promotes running social media accounts, which some apps' own terms may restrict"
        ],
        "agentNotes": [
          "Start with `POST /auth/device/code` and show the person the approval link. Never ask for their email code",
          "Send a new `Idempotency-Key` for each checkout, cancel or input, and reuse it with the same body when you retry",
          "Crop screenshots to `geometry.viewport` and send input coordinates from 0 to 1 within that rectangle",
          "If you get `control_taken`, stop sending input and wait for the person to finish",
          "Check each app's own terms before automating it on the device, since Glasshand leaves that to you"
        ],
        "metrics": {
          "kind": "remote",
          "measured": false
        },
        "reviewCount": 0,
        "avgRating": 0,
        "history": [
          {
            "basis": "public evidence",
            "confidence": "medium",
            "grade": "D",
            "methodology": "1.0",
            "run": "2026-10-01",
            "runLabel": "October 2026 research run",
            "score": 50.8
          }
        ],
        "editorialScores": {
          "ergonomics": 62,
          "maintenance": 35,
          "payments": 35,
          "reliability": 27,
          "schema": 79,
          "security": 60,
          "transparency": 53
        },
        "provenanceScore": 51,
        "checked": {
          "categories": {
            "ergonomics": "2026-11-18",
            "maintenance": "2026-11-18",
            "payments": "2026-11-18",
            "reliability": "2026-11-18",
            "schema": "2026-11-18",
            "security": "2026-11-18",
            "transparency": "2026-11-18"
          },
          "last": "2026-11-18",
          "oldest": "2026-11-18",
          "stale": false,
          "staleAfterDays": 90
        }
      },
      "connect": {
        "http": "curl -X POST https://api.glasshand.com/v1/auth/device/code -H 'Content-Type: application/json' -d '{\"clientName\":\"My agent\"}'"
      },
      "letme": {
        "capability": "https://letme.dev/mobile.control",
        "tool": "https://letme.dev/glasshand"
      },
      "area": "developer",
      "unitPrices": [
        {
          "item": "iPhone SE (3rd generation) rental",
          "unit": "month",
          "usd": 60
        },
        {
          "item": "Mac mini M4 rental",
          "unit": "month",
          "usd": 90
        }
      ],
      "provenance": {
        "legalEntity": "Codapt Inc.",
        "domain": "glasshand.com",
        "domainRegistered": "",
        "endpointOnVendorDomain": true,
        "terms": "https://glasshand.com/terms",
        "privacy": "https://glasshand.com/privacy",
        "statusPage": "",
        "changelog": "",
        "securityTxt": "none",
        "checked": "2026-11-18",
        "notes": [
          "The terms and privacy policy, both effective 22 October 2026, name Codapt Inc., a Delaware corporation, as the operator.",
          "glasshand.com and api.glasshand.com robots.txt allowed every page read on 18 November 2026. glasshand.com was reached from a directory's outbound link and its own pages were read from there.",
          "/.well-known/security.txt returned 404. No status page or changelog is linked from the site.",
          "The domain's registration date wasn't looked up."
        ],
        "score": 51
      },
      "pageJsonUrl": "https://www.anchorterminal.com/tools/glasshand.json",
      "live": {
        "slug": "glasshand",
        "probe": {
          "target": "https://api.glasshand.com/v1",
          "method": "get",
          "lastAt": "2026-11-18T23:13:24.780156329Z",
          "lastOk": true,
          "lastStatus": 401,
          "lastMs": 673,
          "lastNote": "asks for credentials",
          "authRequired": true,
          "uptime24h": 100,
          "uptime30d": 100,
          "p50ms24h": 698,
          "p95ms24h": 1384,
          "samples24h": 16,
          "samples30d": 16,
          "days": [
            {
              "date": "2026-11-18",
              "probes": 16,
              "ok": 16
            }
          ],
          "since": "2026-11-18T21:38:01.809926917Z"
        },
        "updatedAt": "2026-11-18T23:13:24.780156329Z"
      },
      "streak": {
        "days": 0,
        "from": "2026-11-18",
        "sinceWatching": true,
        "best": 0
      }
    },
    "edges": [
      {
        "by": 40,
        "category": "payments",
        "leads": "aura-autopilot",
        "name": "Payments \u0026 pricing",
        "reasons": {
          "a": "Open-source software with paid parts around it, so the pricing lines are read for what a run needs. No payment protocol (0). TypeSafe publishes Jev's price per million input tokens and Aura's plans are public, Indie from $10 a month billed annually (20). Autopilot is free under Apache-2.0, and Aura's two-week trial needs no card (20). No Ramen account is needed with one's own keys, but a person installs the editor plugin and gets a TypeSafe key or signs in to Aura in a browser, so 10 rather than the self-hosted rule's 20 (10). 50 in all. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 50 is 75.",
          "b": "No x402, MPP or L402. Checkout returns a Stripe link (0). Per-device monthly prices published without a login, $60 for an iPhone and $90 for a Mac (30). No free tier or trial (0). An agent can start the device authorisation and prepare a checkout without a key, but a person signs in by email code, approves the agent and pays in a browser (5 of 30). 0 + 30 + 0 + 5 = 35."
        },
        "scores": {
          "a": 75,
          "b": 35
        }
      },
      {
        "by": 14,
        "category": "reliability",
        "leads": "aura-autopilot",
        "name": "Reliability",
        "reasons": {
          "a": "Local-software reading, as a stdio MCP server with engine plugins. Official downloads per engine and OS from GitHub Releases, signed and listed in SHA256SUMS.txt, with Unreal 5.4 to 5.8, Unity 6000.0 or later, Godot 4.7 or later, Windows and macOS and Python 3.12 or later stated. The agent service isn't on a package registry; the plugin installs it, with uv and Bun, on first run (17 of 20). .github/workflows/ci.yml runs ruff and pytest on Linux, Windows and macOS with Python 3.12 and 3.13 plus a Codex end-to-end job, and CONTRIBUTING.md names download and uninstall end-to-end workflows. Whether runs pass wasn't checked, as github.com couldn't be read (15 of 25). Open issues couldn't be read, and the repository is days old (5 of 25). No changelog; releases are tagged v\u003cversion\u003e-\u003ccommit\u003e, the plugin protocol is versioned and the contributing rules say to add optional fields rather than change existing ones (4 of 15). Python package 0.1.0 and the Unreal plugin marked IsBetaVersion (0). 41 in all.",
          "b": "Hosted lines, for the HTTP API. No status page is linked from the site, agents.md or the terms (0), so no incident history could be read (5). No rate limits with numbers were found. agents.md says to back off on 429 (0). Checkout, cancel, resume and each device input take an `Idempotency-Key`, and agents.md says to retry a timeout or 503 with the same key and body and never to replay input blindly after a disconnect. No `Retry-After` is documented (12 of 15). The terms say no particular uptime is guaranteed (0). Registration and paid rentals are open worldwide with live billing, and no beta label was found (10). 0 + 5 + 0 + 12 + 0 + 10 = 27."
        },
        "scores": {
          "a": 41,
          "b": 27
        }
      },
      {
        "by": 8,
        "category": "maintenance",
        "leads": "aura-autopilot",
        "name": "Maintenance \u0026 community",
        "reasons": {
          "a": "The launch post is dated 31 October 2026 and Aura 1.2.3 and 1.2.4 of 30 October 2026 ship Autopilot, so the last change is within 30 days (30). GitHub releases couldn't be read, and the two Aura release-note entries aren't three (0). Issue and pull-request replies couldn't be read; the project is days old, with Discord and support email as other channels (5 of 25). Not in the official MCP registry under any namespace (searched for autopilot and ramenvr) (0). Pinned dependencies, per-platform locks in packaging/, CI on three operating systems, conventional-commit checks and code-owner review (8 of 10). 43 in all.",
          "b": "Closed service with no changelog. The newest first-hand date is the terms and privacy policy effective 22 October 2026, 27 days before this check, read as the launch date, as beam-win's terms date was (30). No dated release entries to count (0). Support by email (support@ and hello@glasshand.com) and a demo booking link. No public tracker or community (5 of 25). No official SDK (0). No public package or CI (0). 30 + 0 + 5 + 0 + 0 = 35."
        },
        "scores": {
          "a": 43,
          "b": 35
        }
      },
      {
        "by": 7,
        "category": "ergonomics",
        "leads": "aura-autopilot",
        "name": "Agent ergonomics",
        "reasons": {
          "a": "MCP reading. 14 tools (15 for 11 to 30), with long descriptions on probe and run_harness and a long INSTRUCTIONS.md the skill says to read in full, with no toolsets (13 of 25). read_file takes offset and limit up to 2,000 lines, list_files returns at most 500, search at most 200 with a glob, frames up to 12 and run_harness takes seconds of play (16 of 20). Errors come back as isError with what to do next (open the editor, update the agent service, call wait), and headless runs exit 0, 1, 2 or 3 (16 of 20). readOnlyHint is set on the seven read tools and destructiveHint false on all, including probe, which the description says can change the game directly. One call runs at a time and a long one answers \"still going\" for `wait` to collect (14 of 20). Two required parameters on start_run with defaults of 10 minutes of exploration and a 30-minute limit; no SDKs, which a playtester doesn't need (10 of 15). 69 in all.",
          "b": "Responses are small, but screenshots are full JPEG frames, and the WebSocket stream holds at most four unacknowledged frames (15 of 25). Rental lists return all items with no pagination or filters, though an account has few rentals (8 of 20). Some error states are named and explained, with no complete error list (12 of 20). An `Idempotency-Key` on every write and every input, with documented retry rules and a conflict on a reused key (20). No SDK. Calls are plain HTTPS with few parameters (7 of 15). 15 + 8 + 12 + 20 + 7 = 62."
        },
        "scores": {
          "a": 69,
          "b": 62
        }
      },
      {
        "by": 7,
        "category": "security",
        "leads": "glasshand",
        "name": "Security \u0026 auth",
        "reasons": {
          "a": "Local-software reading, as tokeneconomizer was read: 30 for what leaves the machine and credentials, 20 approvals and sandboxing, 15 injection, 15 audit, 20 programme. Run sharing is on by default and sends recordings, screenshots, agent traces with the code the agent read and, for MCP runs, the user's own Claude Code or Codex conversation to app.tryaura.dev, after redacting secrets, paths, user and host names and email addresses. Keys sit in plain text in autopilot.env, the file tools never read *.env, and the localhost control endpoint carries a random secret (12). File tools write only the run's folder, notes.md and lib.py, and the user can stop or steer a run from the tab. `autopilot connect codex` sets default_tools_approval_mode to approve, and on Windows and macOS the sandbox is a plain child process without keys but with network and disk (10). sandbox.py names prompt injection from a game's code as the reason for the sandbox, and the frame describer is given only open questions (11). Each run keeps its agent log, harness versions, traces, decisions, report and recording, and the Feed tab shows activity (13). SECURITY.md asks for GitHub private vulnerability reporting; no security.txt on www.tryaura.dev (404), no bounty or certification found, advisories couldn't be read. Releases are signed and immutable, and pull requests need a code owner's approval (7). 53 in all.",
          "b": "OAuth device authorisation with four scopes, 15-minute access tokens, rotating refresh tokens and immediate revocation from the dashboard. No secret in a URL (30). `rentals:read` is a read-only scope, and payment and checkout stay in the person's browser. `devices:control` grants full input with no per-action confirmation, and a person can take control at any time (12 of 20). agents.md treats device screens, web pages, app content and incoming messages as untrusted task data that can't authorise purchases or credential disclosure. No technical control on what an agent does with screen content (10 of 15). Agent authorisation records keep last use, and `control.actor` names the current controller. No audit log of inputs found (6 of 15). No security.txt (404), disclosure policy, bounty or certification. The privacy policy asks for reports to support@glasshand.com (2 of 20). 30 + 12 + 10 + 6 + 2 = 60."
        },
        "scores": {
          "a": 53,
          "b": 60
        }
      },
      {
        "by": 5,
        "category": "schema",
        "leads": "glasshand",
        "name": "Schema \u0026 documentation",
        "reasons": {
          "a": "MCP reading. All 14 tools (start_run, wait, probe, run_harness, snapshot, describe_frames, restart_level, report, frames and five file tools) have JSON Schema inputs, read in autopilot/mcp.py, session.py, external.py and files.py (25). No llms.txt on www.tryaura.dev (404). The agent skill, INSTRUCTIONS.md and KIT.md are Markdown written for the agent, and the README and CONTRIBUTING.md are in the repository (7 of 10). Descriptions say what each tool does and when not to use it (`frames` is for a describer subagent only, `describe_frames` is a fallback), with server instructions (17 of 20). Required fields and an enum on `kind`, but `report.outcome` is a free string with its values only in the description, numeric ranges sit in descriptions rather than the schema, and `probe` takes free-form Python (9 of 15). Prompt examples in the README and skill, plain-language errors in the MCP server and documented exit codes for headless runs (11 of 15). Version numbers and a protocol version, but no public changelog, and GitHub release notes couldn't be read (5 of 15). 74 in all.",
          "b": "A public OpenAPI 3.1 document at api.glasshand.com/openapi.json, version 0.1.0, with 20 operations, linked from agents.md as the API reference (25). agents.md is Markdown written for agents (10). agents.md says when to use each call, how to authenticate, how to retry, how to read the screen and what counts as untrusted data. 15 of 20 operations in the OpenAPI document carry a description (18 of 20). 14 enums and 70 required-field markers, with input types bounded, such as `drag` duration 100 to 3,000 ms and `text` up to 512 characters (13 of 15). Examples in agents.md, and named result errors such as `control_taken`, `control_busy`, `authorization_pending` and an idempotency conflict, but no error reference (8 of 15). Paths are under `/v1`. No changelog found (5 of 15). 25 + 10 + 18 + 13 + 8 + 5 = 79."
        },
        "scores": {
          "a": 74,
          "b": 79
        }
      }
    ],
    "facts": [
      {
        "a": "MCP server",
        "b": "HTTP API",
        "name": "Kind"
      },
      {
        "a": "Ramen",
        "b": "Codapt Inc.",
        "name": "Vendor"
      },
      {
        "a": "no (local only)",
        "b": "https://api.glasshand.com/v1",
        "name": "Hosted endpoint"
      },
      {
        "a": "stdio",
        "b": "HTTP",
        "name": "Transports"
      },
      {
        "a": "API key",
        "b": "OAuth",
        "name": "Auth"
      },
      {
        "a": "Your plan",
        "b": "Paid",
        "name": "Pricing"
      },
      {
        "a": "no",
        "b": "no",
        "name": "x402"
      },
      {
        "a": "Apache-2.0",
        "b": "Proprietary hosted service under the Glasshand Terms of Service",
        "name": "Licence"
      },
      {
        "a": "14",
        "b": "none",
        "name": "Tools exposed"
      },
      {
        "a": "no",
        "b": "no",
        "name": "Read-only variant documented"
      },
      {
        "a": "no",
        "b": "no",
        "name": "llms.txt"
      },
      {
        "a": "none",
        "b": "2026-10-22",
        "name": "Last release"
      },
      {
        "a": "2026-06-25",
        "b": "no date given",
        "name": "Terms last updated"
      },
      {
        "a": "2026-08-27",
        "b": "no date given",
        "name": "Privacy policy last updated"
      },
      {
        "a": "yes, with an opt-out",
        "b": "not found in the text",
        "name": "Customer content may train models"
      },
      {
        "a": "not found in the text",
        "b": "not found in the text",
        "name": "Terms restrict automated access"
      },
      {
        "a": "not found in the text",
        "b": "not found in the text",
        "name": "Terms restrict benchmarking"
      },
      {
        "a": "not found in the text",
        "b": "not found in the text",
        "name": "Terms or service can change without notice"
      },
      {
        "a": "not found in the text",
        "b": "not found in the text",
        "name": "Arbitration or class-action waiver"
      }
    ],
    "faq": [
      {
        "answer": "Aura Autopilot scores 58.7 (C) on agent readiness against Glasshand's 50.8 (D), and leads in 5 of 7 scored categories. Glasshand leads on schema \u0026 documentation and security \u0026 auth.",
        "question": "Which is better for AI agents, Aura Autopilot or Glasshand?"
      },
      {
        "answer": "Aura Autopilot needs an API key. Glasshand uses an OAuth sign-in.",
        "question": "Do Aura Autopilot and Glasshand need an API key?"
      },
      {
        "answer": "Aura Autopilot runs on your own machine, with no hosted endpoint listed. Glasshand has a hosted endpoint at https://api.glasshand.com/v1.",
        "question": "Can an agent call Aura Autopilot and Glasshand without installing anything?"
      },
      {
        "answer": "Aura Autopilot is open source (Apache-2.0). No open-source release is listed for Glasshand.",
        "question": "Are Aura Autopilot and Glasshand open source?"
      }
    ],
    "goodFor": [
      {
        "aheadOn": [
          "Reliability, 41 against 27",
          "Agent ergonomics, 69 against 62",
          "Payments \u0026 pricing, 75 against 35",
          "Maintenance \u0026 community, 43 against 35"
        ],
        "also": [
          "Runs on your own machine",
          "Open source"
        ],
        "goodFor": "Agent-driven playtests of Unreal, Unity or Godot games in the editor or in CI",
        "slug": "aura-autopilot",
        "watchFor": "Run sharing is on by default and sends recordings, screenshots, agent traces with code excerpts and, for MCP runs, the user's Claude Code or Codex conversation to app.tryaura.dev"
      },
      {
        "aheadOn": [
          "Schema \u0026 documentation, 79 against 74",
          "Security \u0026 auth, 60 against 53"
        ],
        "also": [
          "A hosted endpoint, with nothing to install"
        ],
        "goodFor": "Agents that need a real iPhone or Mac, such as testing native apps in Xcode or using iOS-only apps and a real phone number.",
        "slug": "glasshand",
        "watchFor": "No status page, SLA, changelog or security.txt found"
      }
    ],
    "history": {
      "aura-autopilot": [
        {
          "causes": [
            "method"
          ],
          "date": "2026-11-13",
          "delta": 3.2,
          "from": 55.5,
          "fromGrade": "C",
          "to": 58.7,
          "toGrade": "C"
        }
      ],
      "glasshand": null
    },
    "job": {
      "capability": "desktop.control",
      "name": "Desktop app control"
    },
    "others": [
      {
        "json": "https://www.anchorterminal.com/compare/aura-autopilot-vs-autohotkey.json",
        "title": "Aura Autopilot vs AutoHotkey",
        "url": "https://www.anchorterminal.com/compare/aura-autopilot-vs-autohotkey"
      },
      {
        "json": "https://www.anchorterminal.com/compare/aura-autopilot-vs-autoit.json",
        "title": "Aura Autopilot vs AutoIt",
        "url": "https://www.anchorterminal.com/compare/aura-autopilot-vs-autoit"
      },
      {
        "json": "https://www.anchorterminal.com/compare/aura-autopilot-vs-automation-anywhere.json",
        "title": "Aura Autopilot vs Automation Anywhere",
        "url": "https://www.anchorterminal.com/compare/aura-autopilot-vs-automation-anywhere"
      },
      {
        "json": "https://www.anchorterminal.com/compare/aura-autopilot-vs-beam-win.json",
        "title": "Aura Autopilot vs Beam (beam.win)",
        "url": "https://www.anchorterminal.com/compare/aura-autopilot-vs-beam-win"
      },
      {
        "json": "https://www.anchorterminal.com/compare/aura-autopilot-vs-robocorp.json",
        "title": "Aura Autopilot vs Robocorp",
        "url": "https://www.anchorterminal.com/compare/aura-autopilot-vs-robocorp"
      },
      {
        "json": "https://www.anchorterminal.com/compare/aura-autopilot-vs-uipath.json",
        "title": "Aura Autopilot vs UiPath",
        "url": "https://www.anchorterminal.com/compare/aura-autopilot-vs-uipath"
      },
      {
        "json": "https://www.anchorterminal.com/compare/autohotkey-vs-glasshand.json",
        "title": "AutoHotkey vs Glasshand",
        "url": "https://www.anchorterminal.com/compare/autohotkey-vs-glasshand"
      },
      {
        "json": "https://www.anchorterminal.com/compare/autoit-vs-glasshand.json",
        "title": "AutoIt vs Glasshand",
        "url": "https://www.anchorterminal.com/compare/autoit-vs-glasshand"
      },
      {
        "json": "https://www.anchorterminal.com/compare/automation-anywhere-vs-glasshand.json",
        "title": "Automation Anywhere vs Glasshand",
        "url": "https://www.anchorterminal.com/compare/automation-anywhere-vs-glasshand"
      },
      {
        "json": "https://www.anchorterminal.com/compare/beam-win-vs-glasshand.json",
        "title": "Beam (beam.win) vs Glasshand",
        "url": "https://www.anchorterminal.com/compare/beam-win-vs-glasshand"
      },
      {
        "json": "https://www.anchorterminal.com/compare/glasshand-vs-robocorp.json",
        "title": "Glasshand vs Robocorp",
        "url": "https://www.anchorterminal.com/compare/glasshand-vs-robocorp"
      },
      {
        "json": "https://www.anchorterminal.com/compare/glasshand-vs-uipath.json",
        "title": "Glasshand vs UiPath",
        "url": "https://www.anchorterminal.com/compare/glasshand-vs-uipath"
      },
      {
        "json": "https://www.anchorterminal.com/compare/aura-autopilot-vs-robot-framework.json",
        "title": "Aura Autopilot vs Robot Framework",
        "url": "https://www.anchorterminal.com/compare/aura-autopilot-vs-robot-framework"
      },
      {
        "json": "https://www.anchorterminal.com/compare/glasshand-vs-robot-framework.json",
        "title": "Glasshand vs Robot Framework",
        "url": "https://www.anchorterminal.com/compare/glasshand-vs-robot-framework"
      }
    ],
    "panel": null,
    "scores": [
      {
        "aura-autopilot": 41,
        "by": 14,
        "edge": "aura-autopilot",
        "glasshand": 27,
        "key": "reliability",
        "name": "Reliability",
        "weight": 16
      },
      {
        "key": "performance",
        "name": "Performance",
        "pending": true,
        "weight": 10
      },
      {
        "aura-autopilot": 74,
        "by": 5,
        "edge": "glasshand",
        "glasshand": 79,
        "key": "schema",
        "name": "Schema \u0026 documentation",
        "weight": 13
      },
      {
        "aura-autopilot": 69,
        "by": 7,
        "edge": "aura-autopilot",
        "glasshand": 62,
        "key": "ergonomics",
        "name": "Agent ergonomics",
        "weight": 13
      },
      {
        "aura-autopilot": 53,
        "by": 7,
        "edge": "glasshand",
        "glasshand": 60,
        "key": "security",
        "name": "Security \u0026 auth",
        "weight": 14
      },
      {
        "aura-autopilot": 75,
        "by": 40,
        "edge": "aura-autopilot",
        "glasshand": 35,
        "key": "payments",
        "name": "Payments \u0026 pricing",
        "weight": 10
      },
      {
        "key": "tasks",
        "name": "Task success",
        "pending": true,
        "weight": 10
      },
      {
        "aura-autopilot": 43,
        "by": 8,
        "edge": "aura-autopilot",
        "glasshand": 35,
        "key": "maintenance",
        "name": "Maintenance \u0026 community",
        "weight": 7
      },
      {
        "aura-autopilot": 55,
        "by": 3,
        "edge": "aura-autopilot",
        "glasshand": 52,
        "key": "transparency",
        "name": "Transparency \u0026 trust",
        "weight": 7
      }
    ],
    "summary": "Aura Autopilot scores 58.7 (C) on agent readiness against Glasshand's 50.8 (D), and leads in 5 of 7 scored categories. Glasshand leads on schema \u0026 documentation and security \u0026 auth. Both do desktop app control.",
    "tasks": null,
    "tutorials": [],
    "verdicts": {
      "aura-autopilot": "Apache-2.0, with typed MCP tools carrying read-only hints, a separate process for model-written code and headless exit codes for CI. It is days old with no changelog, needs a TypeSafe Jev key or an Aura account, and by default shares recordings, agent traces and the user's own agent conversation with Ramen unless turned off.",
      "glasshand": "Glasshand gives an agent a real iPhone or Mac through a small OpenAPI-described API with OAuth device sign-in, four scopes, short-lived tokens and idempotency keys on every write and input. Rentals are monthly with no trial, and no status page, SLA, changelog or security programme was found."
    }
  },
  "kind": "anchor.page",
  "links": {
    "api": "https://www.anchorterminal.com/api/v1/index.json",
    "html": "https://www.anchorterminal.com/compare/aura-autopilot-vs-glasshand",
    "json": "https://www.anchorterminal.com/compare/aura-autopilot-vs-glasshand.json",
    "llms": "https://www.anchorterminal.com/llms.txt",
    "markdown": "https://www.anchorterminal.com/compare/aura-autopilot-vs-glasshand.md",
    "mcp": "https://www.anchorterminal.com/mcp",
    "slim": "https://www.anchorterminal.com/compare/aura-autopilot-vs-glasshand.min.md"
  },
  "markdown": "Aura Autopilot scores 58.7 (C) on agent readiness against Glasshand's 50.8 (D), and leads in 5 of 7 scored categories. Glasshand leads on schema \u0026 documentation and security \u0026 auth. Both do desktop app control.\n\n- Aura Autopilot: grade C, 58.7/100, rank #840 of 1557. Markdown https://www.anchorterminal.com/tools/aura-autopilot.md · JSON https://www.anchorterminal.com/api/v1/tools/aura-autopilot.json\n- Glasshand: grade D, 50.8/100, rank #1143 of 1557. Markdown https://www.anchorterminal.com/tools/glasshand.md · JSON https://www.anchorterminal.com/api/v1/tools/glasshand.json\n- Best desktop and mobile app automation for AI agents: https://www.anchorterminal.com/best/desktop-automation/index.md\n- All 64 desktop comparisons: https://www.anchorterminal.com/compare/desktop-automation/index.md\n\n## Which one, for what\n\n### Aura Autopilot (C)\n\nGood for: Agent-driven playtests of Unreal, Unity or Godot games in the editor or in CI\n\nAhead on:\n- Reliability, 41 against 27\n- Agent ergonomics, 69 against 62\n- Payments \u0026 pricing, 75 against 35\n- Maintenance \u0026 community, 43 against 35\n\nAlso in its favour:\n- Runs on your own machine\n- Open source\n\nWatch for: Run sharing is on by default and sends recordings, screenshots, agent traces with code excerpts and, for MCP runs, the user's Claude Code or Codex conversation to app.tryaura.dev\n\n### Glasshand (D)\n\nGood for: Agents that need a real iPhone or Mac, such as testing native apps in Xcode or using iOS-only apps and a real phone number.\n\nAhead on:\n- Schema \u0026 documentation, 79 against 74\n- Security \u0026 auth, 60 against 53\n\nAlso in its favour:\n- A hosted endpoint, with nothing to install\n\nWatch for: No status page, SLA, changelog or security.txt found\n\n\n## Why the edge\n\n### Payments \u0026 pricing: Aura Autopilot +40\n\n- Aura Autopilot (75): Open-source software with paid parts around it, so the pricing lines are read for what a run needs. No payment protocol (0). TypeSafe publishes Jev's price per million input tokens and Aura's plans are public, Indie from $10 a month billed annually (20). Autopilot is free under Apache-2.0, and Aura's two-week trial needs no card (20). No Ramen account is needed with one's own keys, but a person installs the editor plugin and gets a TypeSafe key or signs in to Aura in a browser, so 10 rather than the self-hosted rule's 20 (10). 50 in all. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 50 is 75.\n- Glasshand (35): No x402, MPP or L402. Checkout returns a Stripe link (0). Per-device monthly prices published without a login, $60 for an iPhone and $90 for a Mac (30). No free tier or trial (0). An agent can start the device authorisation and prepare a checkout without a key, but a person signs in by email code, approves the agent and pays in a browser (5 of 30). 0 + 30 + 0 + 5 = 35.\n\n### Reliability: Aura Autopilot +14\n\n- Aura Autopilot (41): Local-software reading, as a stdio MCP server with engine plugins. Official downloads per engine and OS from GitHub Releases, signed and listed in SHA256SUMS.txt, with Unreal 5.4 to 5.8, Unity 6000.0 or later, Godot 4.7 or later, Windows and macOS and Python 3.12 or later stated. The agent service isn't on a package registry; the plugin installs it, with uv and Bun, on first run (17 of 20). .github/workflows/ci.yml runs ruff and pytest on Linux, Windows and macOS with Python 3.12 and 3.13 plus a Codex end-to-end job, and CONTRIBUTING.md names download and uninstall end-to-end workflows. Whether runs pass wasn't checked, as github.com couldn't be read (15 of 25). Open issues couldn't be read, and the repository is days old (5 of 25). No changelog; releases are tagged v\u003cversion\u003e-\u003ccommit\u003e, the plugin protocol is versioned and the contributing rules say to add optional fields rather than change existing ones (4 of 15). Python package 0.1.0 and the Unreal plugin marked IsBetaVersion (0). 41 in all.\n- Glasshand (27): Hosted lines, for the HTTP API. No status page is linked from the site, agents.md or the terms (0), so no incident history could be read (5). No rate limits with numbers were found. agents.md says to back off on 429 (0). Checkout, cancel, resume and each device input take an `Idempotency-Key`, and agents.md says to retry a timeout or 503 with the same key and body and never to replay input blindly after a disconnect. No `Retry-After` is documented (12 of 15). The terms say no particular uptime is guaranteed (0). Registration and paid rentals are open worldwide with live billing, and no beta label was found (10). 0 + 5 + 0 + 12 + 0 + 10 = 27.\n\n### Maintenance \u0026 community: Aura Autopilot +8\n\n- Aura Autopilot (43): The launch post is dated 31 October 2026 and Aura 1.2.3 and 1.2.4 of 30 October 2026 ship Autopilot, so the last change is within 30 days (30). GitHub releases couldn't be read, and the two Aura release-note entries aren't three (0). Issue and pull-request replies couldn't be read; the project is days old, with Discord and support email as other channels (5 of 25). Not in the official MCP registry under any namespace (searched for autopilot and ramenvr) (0). Pinned dependencies, per-platform locks in packaging/, CI on three operating systems, conventional-commit checks and code-owner review (8 of 10). 43 in all.\n- Glasshand (35): Closed service with no changelog. The newest first-hand date is the terms and privacy policy effective 22 October 2026, 27 days before this check, read as the launch date, as beam-win's terms date was (30). No dated release entries to count (0). Support by email (support@ and hello@glasshand.com) and a demo booking link. No public tracker or community (5 of 25). No official SDK (0). No public package or CI (0). 30 + 0 + 5 + 0 + 0 = 35.\n\n### Agent ergonomics: Aura Autopilot +7\n\n- Aura Autopilot (69): MCP reading. 14 tools (15 for 11 to 30), with long descriptions on probe and run_harness and a long INSTRUCTIONS.md the skill says to read in full, with no toolsets (13 of 25). read_file takes offset and limit up to 2,000 lines, list_files returns at most 500, search at most 200 with a glob, frames up to 12 and run_harness takes seconds of play (16 of 20). Errors come back as isError with what to do next (open the editor, update the agent service, call wait), and headless runs exit 0, 1, 2 or 3 (16 of 20). readOnlyHint is set on the seven read tools and destructiveHint false on all, including probe, which the description says can change the game directly. One call runs at a time and a long one answers \"still going\" for `wait` to collect (14 of 20). Two required parameters on start_run with defaults of 10 minutes of exploration and a 30-minute limit; no SDKs, which a playtester doesn't need (10 of 15). 69 in all.\n- Glasshand (62): Responses are small, but screenshots are full JPEG frames, and the WebSocket stream holds at most four unacknowledged frames (15 of 25). Rental lists return all items with no pagination or filters, though an account has few rentals (8 of 20). Some error states are named and explained, with no complete error list (12 of 20). An `Idempotency-Key` on every write and every input, with documented retry rules and a conflict on a reused key (20). No SDK. Calls are plain HTTPS with few parameters (7 of 15). 15 + 8 + 12 + 20 + 7 = 62.\n\n### Security \u0026 auth: Glasshand +7\n\n- Aura Autopilot (53): Local-software reading, as tokeneconomizer was read: 30 for what leaves the machine and credentials, 20 approvals and sandboxing, 15 injection, 15 audit, 20 programme. Run sharing is on by default and sends recordings, screenshots, agent traces with the code the agent read and, for MCP runs, the user's own Claude Code or Codex conversation to app.tryaura.dev, after redacting secrets, paths, user and host names and email addresses. Keys sit in plain text in autopilot.env, the file tools never read *.env, and the localhost control endpoint carries a random secret (12). File tools write only the run's folder, notes.md and lib.py, and the user can stop or steer a run from the tab. `autopilot connect codex` sets default_tools_approval_mode to approve, and on Windows and macOS the sandbox is a plain child process without keys but with network and disk (10). sandbox.py names prompt injection from a game's code as the reason for the sandbox, and the frame describer is given only open questions (11). Each run keeps its agent log, harness versions, traces, decisions, report and recording, and the Feed tab shows activity (13). SECURITY.md asks for GitHub private vulnerability reporting; no security.txt on www.tryaura.dev (404), no bounty or certification found, advisories couldn't be read. Releases are signed and immutable, and pull requests need a code owner's approval (7). 53 in all.\n- Glasshand (60): OAuth device authorisation with four scopes, 15-minute access tokens, rotating refresh tokens and immediate revocation from the dashboard. No secret in a URL (30). `rentals:read` is a read-only scope, and payment and checkout stay in the person's browser. `devices:control` grants full input with no per-action confirmation, and a person can take control at any time (12 of 20). agents.md treats device screens, web pages, app content and incoming messages as untrusted task data that can't authorise purchases or credential disclosure. No technical control on what an agent does with screen content (10 of 15). Agent authorisation records keep last use, and `control.actor` names the current controller. No audit log of inputs found (6 of 15). No security.txt (404), disclosure policy, bounty or certification. The privacy policy asks for reports to support@glasshand.com (2 of 20). 30 + 12 + 10 + 6 + 2 = 60.\n\n### Schema \u0026 documentation: Glasshand +5\n\n- Aura Autopilot (74): MCP reading. All 14 tools (start_run, wait, probe, run_harness, snapshot, describe_frames, restart_level, report, frames and five file tools) have JSON Schema inputs, read in autopilot/mcp.py, session.py, external.py and files.py (25). No llms.txt on www.tryaura.dev (404). The agent skill, INSTRUCTIONS.md and KIT.md are Markdown written for the agent, and the README and CONTRIBUTING.md are in the repository (7 of 10). Descriptions say what each tool does and when not to use it (`frames` is for a describer subagent only, `describe_frames` is a fallback), with server instructions (17 of 20). Required fields and an enum on `kind`, but `report.outcome` is a free string with its values only in the description, numeric ranges sit in descriptions rather than the schema, and `probe` takes free-form Python (9 of 15). Prompt examples in the README and skill, plain-language errors in the MCP server and documented exit codes for headless runs (11 of 15). Version numbers and a protocol version, but no public changelog, and GitHub release notes couldn't be read (5 of 15). 74 in all.\n- Glasshand (79): A public OpenAPI 3.1 document at api.glasshand.com/openapi.json, version 0.1.0, with 20 operations, linked from agents.md as the API reference (25). agents.md is Markdown written for agents (10). agents.md says when to use each call, how to authenticate, how to retry, how to read the screen and what counts as untrusted data. 15 of 20 operations in the OpenAPI document carry a description (18 of 20). 14 enums and 70 required-field markers, with input types bounded, such as `drag` duration 100 to 3,000 ms and `text` up to 512 characters (13 of 15). Examples in agents.md, and named result errors such as `control_taken`, `control_busy`, `authorization_pending` and an idempotency conflict, but no error reference (8 of 15). Paths are under `/v1`. No changelog found (5 of 15). 25 + 10 + 18 + 13 + 8 + 5 = 79.\n\n## Score by category\n\n| Category | Weight | Aura Autopilot | Glasshand | Edge |\n| --- | --- | --- | --- | --- |\n| Reliability | 16% | 41 | 27 | Aura Autopilot +14 |\n| Performance | 10%, pending | pending | pending | not scored in this run |\n| Schema \u0026 documentation | 13% | 74 | 79 | Glasshand +5 |\n| Agent ergonomics | 13% | 69 | 62 | Aura Autopilot +7 |\n| Security \u0026 auth | 14% | 53 | 60 | Glasshand +7 |\n| Payments \u0026 pricing | 10% | 75 | 35 | Aura Autopilot +40 |\n| Task success | 10%, pending | pending | pending | not scored in this run |\n| Maintenance \u0026 community | 7% | 43 | 35 | Aura Autopilot +8 |\n| Transparency \u0026 trust | 7% | 55 | 52 | Aura Autopilot +3 |\n| Negative events | ≤15 | 0 | 0 | |\n| **Total** | | **58.7 · C** | **50.8 · D** | |\n\n## Facts side by side\n\n| Fact | Aura Autopilot | Glasshand |\n| --- | --- | --- |\n| Kind | MCP server | HTTP API |\n| Vendor | Ramen | Codapt Inc. |\n| Hosted endpoint | no (local only) | `https://api.glasshand.com/v1` |\n| Transports | stdio | HTTP |\n| Auth | API key | OAuth |\n| Pricing | Your plan | Paid |\n| x402 | no | no |\n| Licence | Apache-2.0 | Proprietary hosted service under the Glasshand Terms of Service |\n| Tools exposed | 14 | none |\n| Read-only variant documented | no | no |\n| llms.txt | no | no |\n| Last release | none | 2026-10-22 |\n| Terms last updated | 2026-06-25 | no date given |\n| Privacy policy last updated | 2026-08-27 | no date given |\n| Customer content may train models | yes, with an opt-out | not found in the text |\n| Terms restrict automated access | not found in the text | not found in the text |\n| Terms restrict benchmarking | not found in the text | not found in the text |\n| Terms or service can change without notice | not found in the text | not found in the text |\n| Arbitration or class-action waiver | not found in the text | not found in the text |\n\n## How the scores have moved\n\n### Aura Autopilot\n\n- +3.2 on 13 November 2026: checklist\n\n### Glasshand\n\nNo score change since it was listed.\n\n## Verdicts\n\n**Aura Autopilot.** Apache-2.0, with typed MCP tools carrying read-only hints, a separate process for model-written code and headless exit codes for CI. It is days old with no changelog, needs a TypeSafe Jev key or an Aura account, and by default shares recordings, agent traces and the user's own agent conversation with Ramen unless turned off.\n\n**Glasshand.** Glasshand gives an agent a real iPhone or Mac through a small OpenAPI-described API with OAuth device sign-in, four scopes, short-lived tokens and idempotency keys on every write and input. Rentals are monthly with no trial, and no status page, SLA, changelog or security programme was found.\n\n## Before you call either\n\n### Aura Autopilot\n\n1. Check the game's project is open in the editor with the Autopilot tab connected before calling start_run; otherwise it answers that the agent service isn't running.\n2. When a call answers \"still going\", call `wait` until it returns; don't start another tool call meanwhile.\n3. Pass the user's request in their own words with kind verify, goal or explore, and read the INSTRUCTIONS.md that start_run names in parts with Autopilot's read_file.\n4. Tell the user that run data is shared by default and how to turn it off (the Settings page, AUTOPILOT_SHARE=0 or -PlaytestShare=0) before the first run.\n5. Never ask the user to paste a TypeSafe or provider key into the conversation; keys belong in autopilot.env or the Autopilot tab.\n\n### Glasshand\n\n1. Start with `POST /auth/device/code` and show the person the approval link. Never ask for their email code\n2. Send a new `Idempotency-Key` for each checkout, cancel or input, and reuse it with the same body when you retry\n3. Crop screenshots to `geometry.viewport` and send input coordinates from 0 to 1 within that rectangle\n4. If you get `control_taken`, stop sending input and wait for the person to finish\n5. Check each app's own terms before automating it on the device, since Glasshand leaves that to you\n\n## Questions\n\n### Which is better for AI agents, Aura Autopilot or Glasshand?\n\nAura Autopilot scores 58.7 (C) on agent readiness against Glasshand's 50.8 (D), and leads in 5 of 7 scored categories. Glasshand leads on schema \u0026 documentation and security \u0026 auth.\n\n### Do Aura Autopilot and Glasshand need an API key?\n\nAura Autopilot needs an API key. Glasshand uses an OAuth sign-in.\n\n### Can an agent call Aura Autopilot and Glasshand without installing anything?\n\nAura Autopilot runs on your own machine, with no hosted endpoint listed. Glasshand has a hosted endpoint at https://api.glasshand.com/v1.\n\n### Are Aura Autopilot and Glasshand open source?\n\nAura Autopilot is open source (Apache-2.0). No open-source release is listed for Glasshand.\n\n\n## For agents\n\n- This comparison as JSON: https://www.anchorterminal.com/compare/aura-autopilot-vs-glasshand.json, and with the fewest tokens: https://www.anchorterminal.com/compare/aura-autopilot-vs-glasshand.min.md\n- Over MCP at https://www.anchorterminal.com/mcp (no key): `compare_tools {\"a\": \"aura-autopilot\", \"b\": \"glasshand\"}`. From a terminal: `anchor compare aura-autopilot glasshand`\n- Each listing in full: https://www.anchorterminal.com/api/v1/tools/aura-autopilot.json and https://www.anchorterminal.com/api/v1/tools/glasshand.json\n\n## Other comparisons with Aura Autopilot or Glasshand\n\n- [Aura Autopilot vs AutoHotkey](https://www.anchorterminal.com/compare/aura-autopilot-vs-autohotkey.md)\n- [Aura Autopilot vs AutoIt](https://www.anchorterminal.com/compare/aura-autopilot-vs-autoit.md)\n- [Aura Autopilot vs Automation Anywhere](https://www.anchorterminal.com/compare/aura-autopilot-vs-automation-anywhere.md)\n- [Aura Autopilot vs Beam (beam.win)](https://www.anchorterminal.com/compare/aura-autopilot-vs-beam-win.md)\n- [Aura Autopilot vs Robocorp](https://www.anchorterminal.com/compare/aura-autopilot-vs-robocorp.md)\n- [Aura Autopilot vs UiPath](https://www.anchorterminal.com/compare/aura-autopilot-vs-uipath.md)\n- [AutoHotkey vs Glasshand](https://www.anchorterminal.com/compare/autohotkey-vs-glasshand.md)\n- [AutoIt vs Glasshand](https://www.anchorterminal.com/compare/autoit-vs-glasshand.md)\n- [Automation Anywhere vs Glasshand](https://www.anchorterminal.com/compare/automation-anywhere-vs-glasshand.md)\n- [Beam (beam.win) vs Glasshand](https://www.anchorterminal.com/compare/beam-win-vs-glasshand.md)\n- [Glasshand vs Robocorp](https://www.anchorterminal.com/compare/glasshand-vs-robocorp.md)\n- [Glasshand vs UiPath](https://www.anchorterminal.com/compare/glasshand-vs-uipath.md)\n- [Aura Autopilot vs Robot Framework](https://www.anchorterminal.com/compare/aura-autopilot-vs-robot-framework.md)\n- [Glasshand vs Robot Framework](https://www.anchorterminal.com/compare/glasshand-vs-robot-framework.md)\n",
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "attributionRequired": true,
    "credit": "CC BY 4.0, credit Anchor Terminal and link the page or API address you took it from (https://www.anchorterminal.com/terms/#data)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-11-18",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "1.0",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run",
    "terms": "https://www.anchorterminal.com/terms/#data"
  },
  "page": {
    "breadcrumbs": [
      {
        "name": "Home",
        "url": "https://www.anchorterminal.com/"
      },
      {
        "name": "Compare",
        "url": "https://www.anchorterminal.com/compare/"
      },
      {
        "name": "Aura Autopilot vs Glasshand",
        "url": ""
      }
    ],
    "description": "Aura Autopilot scores 58.7 (C) to Glasshand's 50.8 (D) for desktop app control. Prices, MCP, x402, uptime and agent notes side by side.",
    "facts": [
      "Aura Autopilot C 58.7",
      "Glasshand D 50.8",
      "scores"
    ],
    "h1": "Aura Autopilot vs Glasshand",
    "image": "https://www.anchorterminal.com/assets/og/default.png",
    "path": "/compare/aura-autopilot-vs-glasshand",
    "published": "2026-10-01",
    "section": "tools",
    "title": "Aura Autopilot vs Glasshand for AI agents in 2026: scores and prices",
    "toc": null,
    "updated": "2026-11-18",
    "url": "https://www.anchorterminal.com/compare/aura-autopilot-vs-glasshand"
  },
  "tokens": {
    "markdown": 4750,
    "slim": 980
  },
  "version": 1
}
