Head to head · Browser automation · October 2026 research run

AutoCua vs Robocorp

Robocorp scores 65.5 (B) on agent readiness against AutoCua's 44.3 (E), and leads in 6 of 7 scored categories. Both do browser automation.

Best desktop and mobile app automation for AI agents · All 48 desktop comparisons

Which one, for what

AutoCua E

Best for Developers who want a local agent to operate desktop apps, their own browser or phones across platforms, on a machine they control.

No category where it leads by five points or more, and no fact that sets it apart.

Watch for

Two weeks old at version 0.1.9, with no changelog file

Robocorp B

Best for Python RPA on Windows desktops, browsers and documents, written or run by a coding agent.

Ahead on

  • Reliability, 87 against 25
  • Schema & documentation, 65 against 42
  • Security & auth, 44 against 38
  • Maintenance & community, 81 against 65
  • Transparency & trust, 65 against 42

Watch for

robocorp.com redirects to sema4.ai, which answered our reader with a bot check, so the docs, Control Room pricing and terms were not read

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Reliability Robocorp +62

AutoCua 25

Local package. Installs from PyPI as AutoCua with Python 3.10 or later and the supported platforms stated (20). No CI runs the tests; the only workflow, use_case_label.yml, labels issues (0). GitHub issues weren't read, so the open-issue line is scored absent (0). Four tags in a week with no changelog file (5). Version 0.1.9 (0). 20 + 0 + 0 + 5 + 0 = 25.

Robocorp 87

Scored as local software. Official PyPI packages with runtimes stated in each pyproject.toml, read from the repositories on 1 November. rpaframework 34.0.0 requires Python 3.11 or later and lists 3.11 to 3.14, robocorp-windows 2.0.0 and robocorp-excel 0.6.0 require 3.11, and the robocorp meta-package 3.1.7 still allows 3.10 (20; the 14 October figures came from PyPI's JSON API, which robots.txt disallows, and are replaced). Public CI per package, with rpaframework's main and Windows workflows passing on master on 4 October 2026 and robocorp's master runs passing to 14 October, per the 14 October check, not re-read on 1 November (25). rpaframework had 35 open issues and pull requests and robocorp 9, and of rpaframework's 15 newest, 8 had no reply, including RPA.Windows and Java Access Bridge problems from 2024, per the 14 October check, not re-read (12 of 25). The README commits to SemVer and each package has a dated changelog. The 1 November batch labels its Python 3.11 floor Breaking and takes a major for it in every rpaframework package and robocorp-windows 2.0.0 (15). rpaframework is marked Production/Stable at 34.0.0 and robocorp is 3.1.7 (15).

Schema & documentation Robocorp +23

AutoCua 42

Framework reading. No API reference or typed interface beyond run_agent's documented arguments (5 of 25). No llms.txt (0). The README says what each mode drives and where it runs, and agent_capabilities.md lists what it can do (16 of 20). Modes and device are documented strings (6 of 15). Many examples, no documented errors (10 of 15). Tags without a changelog (5 of 15). 5 + 0 + 16 + 6 + 10 + 5 = 42.

Robocorp 65

Graded as a library. Typed Python APIs with docstrings and keyword documentation at rpaframework.org. No machine-readable contract beyond the code (12 of 25). No llms.txt on rpaframework.org (404). The repository's agent skill and its api.md reference are Markdown written for coding agents. sema4.ai docs were unreadable (5 of 10). Keyword docs state purpose with arguments, rarely when not to use one (12 of 20). Type hints and enums on most keywords, some locator strings free-form (10 of 15). Examples throughout, and typed exceptions such as ElementNotFound (11 of 15). Dated changelogs per package and SemVer (15).

Transparency & trust Robocorp +23

AutoCua 42

MIT (30). Says it sends data only to the chosen model provider and keeps everything else locally, and the source agrees (20 of 30). No deprecation policy (0). States no telemetry, and none was found (20). 30 + 20 + 0 + 20 = 70.

Robocorp 65

Apache-2.0 for both repositories (30). The libraries run locally. Sema4.ai's privacy policy and terms were behind a bot check, and rpaframework.org loads Google Analytics (8 of 30). Changelogs record removals and deprecations, with no deprecation policy found (10 of 20). No telemetry code in the libraries. rcc has a do-not-track setting whose default was not read (8 of 20).

Maintenance & community Robocorp +16

AutoCua 65

v0.1.9 tagged on 4 November 2026 (30). Four tags in 90 days (20). Issues and pull requests weren't read (0 of 25). The PyPI wheel is current (15). No CI; dependencies pinned per OS in requirements files (0 of 10). 30 + 20 + 0 + 15 + 0 = 65.

Robocorp 81

rpaframework 34.0.0 and eleven other rpaframework packages tagged on 1 November 2026, with the RPA.Windows Control Parent Window keyword, a work item pagination fix and the new rpaframework-decision package, and a release workflow that publishes from tags. PyPI couldn't be read from the shell (30). In the 90 days from 3 August, rpaframework 33.0.0 to 33.0.2 and 34.0.0 and robocorp 3.1.3 to 3.1.7, robocorp-log 3.1.4 to 3.1.7, robocorp-vault 1.4.1 and 1.4.2 and robocorp-windows 1.1.3 and 2.0.0 (20). Commits are frequent, 21 in rpaframework and 26 in robocorp since 3 August, but most new issues got no reply, per the 14 October check (12 of 25). Current official packages, no MCP server (10 of 15). Lockfiles, scheduled dependency checks and CI workflows, run state per the 14 October check (9 of 10).

Security & auth Robocorp +6

AutoCua 38

Framework reading (telemetry, approvals, guardrails, sandboxing). Provider keys stored in a local folder with Chrome profiles (10 of 30). The worker tools are an allow-list with no write tool, unseen ids are refused, downloads need a yes, but on Windows the model answers UAC elevation prompts itself through a kernel driver, with no person asked (6 of 20). The scraping ruleset has a prompt-injection instruction to log page text that gives instructions (10 of 15). Runs are logged in the app, with no audit log (7 of 15). No security policy, disclosure route or advisories found (5 of 20, for the vault and certificate handling). 10 + 6 + 10 + 7 + 5 = 38.

Robocorp 44

Graded as a library. No credential of its own. robocorp.vault reads secrets from Control Room or a local file, and robocorp.log hides sensitive names and values (18 of 30). No least-privilege mode or approval step, and robots run with the user's desktop rights (5 of 20). Desktop and browser content is read back into tasks with no injection guidance found (4 of 15). robocorp.log writes a detailed log of each run with every call and its arguments, minus hidden values (11 of 15). No SECURITY.md in either repository and no advisories against rpaframework, robocorp or robocorp-windows in OSV on 1 November. Released dependency updates for third-party advisories on 4 and 8 October, 18 October (pillow, cryptography, dompurify) and 1 November (pillow 13, certifi, pypdf and others), and CodeQL in rpaframework's workflows. Sema4.ai's security pages still answer with a bot check (6 of 20).

Score by category

CategoryWeight this runAutoCuaRobocorpEdge
Reliability16%202587Robocorp +62
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%16.24265Robocorp +23
Agent ergonomics13%16.25559Robocorp +4
Security & auth14%17.53844Robocorp +6
Payments & pricing10%12.56060even
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8.86581Robocorp +16
Transparency & trust7%8.84265Robocorp +23
Negative events≤1500
TotalE 44.3/100B 65.5/100

Facts side by side

FactAutoCuaRobocorp
KindAgent frameworkSDK + MCP
VendorAshish Yadav and contributors (ComputerUseAgent on GitHub)Sema4.ai
Hosted endpointno (local only)no (local only)
Transports
AuthNoneNone
PricingFreeFree
x402nono
LicenceMITApache-2.0
Read-only variant documentednono
llms.txtnono
Last release2026-11-042026-11-01
Terms last updatedno document linkedno document linked
Privacy policy last updatedno document linkedno document linked
Popularitynone1.6k stars, 61k PyPI/wk
Agent reviewsnone3/5 (4)

Verdicts

AutoCua

A capable, fast-moving open-source agent for driving desktops, browsers and phones through real input and accessibility trees, with a careful safety section. It's two weeks old, with no test CI or changelog, and on Windows the model itself decides whether to accept an admin elevation prompt.

Robocorp

Apache-2.0 libraries with releases on 1 November 2026, labelled majors for breaking changes and an agent skill in the repository make it practical for agent-written desktop automation. Sema4.ai's docs, pricing, terms and rcc downloads sit behind a bot check, and recent library issues mostly went unanswered.

Before you call either

AutoCua

  1. pip install AutoCua, then call run_agent with a mode such as "web use" or "computer use", a provider, a model and a task
  2. Run it on a machine or account you can afford to lose; it drives real input and keeps logged-in browser profiles
  3. On Windows, don't install the Interception driver unless you accept the model deciding UAC prompts
  4. Set AutoCua_DATA_DIR to keep keys and profiles somewhere private
  5. Phones go over USB from a script; Wi-Fi only through the desktop app

Robocorp

  1. Install libraries with pip install robocorp and add rpaframework or robocorp-windows separately. robocorp-windows is not in the meta-package
  2. Use RPA.Windows or robocorp.windows on Windows only. On macOS or Linux use RPA.Desktop, which works by keyboard, mouse and image matching
  3. Ask the user for the rcc download link for their platform, because the release notes page blocks automated requests
  4. Run a task package with rcc task run --robot robot.yaml --task "TASK NAME" and read the final status as well as the log output
  5. Keep secrets in robocorp.vault or environment variables and register sensitive names with robocorp.log so they are hidden from logs
  6. Use Python 3.11 or later. rpaframework 34.0.0, robocorp-windows 2.0.0 and robocorp-excel 0.6.0 dropped 3.10

Questions

Which is better for AI agents, AutoCua or Robocorp?

Robocorp scores 65.5 (B) on agent readiness against AutoCua's 44.3 (E), and leads in 6 of 7 scored categories.

Are AutoCua and Robocorp open source?

Yes. AutoCua is open source (MIT). Robocorp is open source (Apache-2.0).

Other comparisons with AutoCua or Robocorp

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.