Is Ghostwork worth it for a solo macOS power user?

If you already repeat reversible, browser-heavy flows — LinkedIn search to CRM export, inbox triage — yes. You get the timeline view, natural-language search and app usage analytics for $0 that Screenpipe's desktop app charges $25/month for, plus observation-based automation on top. The trade is alpha code (0.1.x, 20 commits at the repo scrape), a manual git clone plus npm install, and an OpenRouter key for the Claude vision fallback. Take it if you live in a terminal; skip it if you want a supported product.

Does Ghostwork integrate with Screenpipe and OpenRouter?

Yes, both are core. Ghostwork is a GUI plus automation agent built on the free Screenpipe CLI — L2 episodic memory ingests clicks, keystrokes and app switches every 2 minutes from Screenpipe's input stream. OpenRouter supplies the LLM key used for the nightly NREM/REM/GC consolidation that promotes patterns into semantic rules and compiled skills. Anthropic appears too, via the Claude vision fallback for browsers and AX-empty apps. No other integrations are documented in the repo.

How does Ghostwork compare to the paid Screenpipe desktop app?

Screenpipe's desktop app moved to $25/month. Ghostwork works with the free Screenpipe CLI and gives you a full GUI layer covering the timeline view, natural-language search and app usage analytics at $0 under GPL-3.0, then adds macOS workflow automation, learning from observation and autonomous execution — three rows the vendor's own comparison table marks as absent from the $25/month desktop app. The difference: Ghostwork is source you build and run yourself, Screenpipe desktop is a maintained app.

What's the cheapest Ghostwork tier? Is Ghostwork free?

Ghostwork has exactly one tier: Free (open source), $0, GPL-3.0, no seats and no subscription. That covers the timeline view, natural-language search, app usage analytics, the four memory layers and the automation agent. Your real cost shifts elsewhere — you pay for the OpenRouter (or Anthropic) key that powers the Claude vision fallback and the nightly consolidation, plus your own setup and maintenance time.

What are Ghostwork's biggest limitations?

Three matter most. It is 0.1.x-alpha with 20 commits and 12 open issues at the repo scrape, so expect bugs. Setup is manual — Node.js 20+, Screenpipe CLI, npm install and an OpenRouter key. And it is macOS 12+ only, with no cloud sync, shared memory or admin controls, plus rollback that only covers undoable actions. The app itself is free, but browser automation routes through a paid vision fallback on your key.

How long does Ghostwork take to set up?

A developer should budget 30-60 minutes: clone the repo, npm install, confirm Node.js 20+, install and run the Screenpipe CLI, and add an OpenRouter key. Non-developers should plan an afternoon and expect to debug. Usefully, nothing is learned at install — the first L3 semantic rules only appear after a nightly consolidation cycle, and rules need 5 accepts with fewer than 2 rejections in the last 10 runs to graduate to Autonomous.

Can Ghostwork automate tasks without prompts or integrations?

Yes — that is its core claim. L2 episodic memory ingests raw clicks, keystrokes and app switches every 2 minutes; the nightly NREM pass promotes repeated patterns into L3 rules like 'WHEN on LinkedIn search → DO export to CRM', and REM compiles rules with 3+ observations into L4 executable skills with DOM/AX locators. You never write a prompt. Execution is AX-first at roughly 95% accuracy on native apps, with Claude vision for browsers and zero-token deterministic replay for compiled skills.

Is Ghostwork safe to let run autonomously?

Its design assumes it isn't, at first. Every new rule starts Supervised — it executes immediately, shows a HUD notification and stays undoable with Cmd+Z. Graduation to Autonomous requires 5 accepts and fewer than 2 rejections in the last 10 runs. Anything externally visible (send email, submit form, post) always requires one-tap approval regardless of tier, and a failed multi-step sequence fires Cmd+Z in reverse order before surfacing the error. Rollback only covers reversible steps, so avoid irreversible workflows.

Does Ghostwork replace a commercial RPA tool?

For a solo macOS user doing reversible browser flows, it can — observation-based learning removes the selector maintenance, and compiled skills replay deterministically at zero tokens. It does not replace RPA for teams: there's no Windows or Linux build, no cloud sync, shared memory or admin controls, and no vendor support contract behind 0.1.x-alpha code with a handful of contributors. Look elsewhere if you need an SLA or a heterogeneous fleet.

Is Ghostwork good for automating browser workflows?

Yes, with a cost caveat. Compiled skills replay recorded step sequences deterministically at zero tokens, but browsers and AX-empty apps fall through to the pixel-level Claude vision fallback, which bills against your own OpenRouter or Anthropic key. So high-volume browser automation has a real per-call cost even though the app is $0 — the repo's own note puts AX accuracy at roughly 95% on native apps, which is where the cheap path lives.

Is ghostwork still active in 2026?

ghostwork is active in 2026 but worth monitoring — liveness 69/100. Its weakest signal is what users say about it.

Browser & Computer-Use Agents

ghostwork

Ghostwork is a free, open-source macOS screen-recording GUI and workflow automation agent built on the Screenpipe CLI.

69/100MonitorFreeFree

Ghostwork is the most ambitious free GUI on top of Screenpipe and the only option here that turns what it records into automation — the four memory layers and supervised-to-autonomous tiering are real engineering, not a prompt wrapper. Privacy-focused Mac users comfortable in a terminal will get genuine leverage from it. Everyone else should wait: 0.1.x-alpha code, 20 commits, manual setup, and your own OpenRouter key are not a support contract.

Verified 1d ago · liveness 69/100 · cite: rightaichoice.com/tools/ghostwork

Best for

macOS power users who want a free GUI over the Screenpipe CLI
Developers comfortable with git clone, npm install and managing an OpenRouter key
Privacy-first users who need screen history to stay on-device
People repeating the same multi-step browser flows daily

Not ideal for

Windows or Linux users — macOS 12+ only
Buyers who need a supported product with a vendor to escalate to
Teams needing cloud sync, shared memory or admin controls

Visit Website

IntermediateExpect 30-60 minutes before first value for a developer: clone the repo, npm install, confirm Node.js 20+, install and run the Screenpipe CLI, and add an OpenRouter key. Non-developers should budget an afternoon, and should expect to debug. Nothing meaningful is learned until the agent has recorded a few working sessions — the first L3 semantic rules only appear after a nightly consolidationDesktopNo public APIVerified 1d ago

Pricing

Free

FreeFree tier4 hidden costs

Learning curve

Intermediate

Expect 30-60 minutes before first value for a developer: clone the repo, npm install, confirm Node.js 20+, install and run the Screenpipe CLI, and add an OpenRouter key. Non-developers should budget an afternoon, and should expect to debug. Nothing meaningful is learned until the agent has recorded a few working sessions — the first L3 semantic rules only appear after a nightly consolidation

Runs on

Desktop

No public API · 4 integrations

Who it's for

Solo recruiter running LinkedIn sourcing all dayPrivacy-conscious developerOps generalist tired of form filling

Live sentiment

Is ghostwork actually worth it?

We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.

Honest verdict, not marketing
Real pros & cons from real users
Attributed quotes with receipts

Run a free scan

3 free scans · no card needed

Skip it if

Skip Ghostwork if you need a supported, turnkey Mac app you can install and forget — this is 0.1.x-alpha software you build from source, wire to an OpenRouter key and debug yourself.

The 30-second take

Biggest gripe

The app is free, but browser and AX-empty workflows route through the Claude vision fallback, so each fallback call bills against your own OpenRouter or Anthropic key.

Price reality

Ghostwork is $0 under GPL-3.0 with no seats, no subscription and no per-user fees — it undercuts Screenpipe's desktop app at $25/month outright. The catch is that your real spend shifts to an OpenRouter or Anthropic key for the Claude vision fallback and nightly consolidation, plus your own setup time. For a solo Mac user doing reversible browser flows, that is far cheaper than $25/month; for a team that needs SSO, shared memory or a support contract, the paid alternatives are still the right

In short

ghostwork — Ghostwork is a free, open-source macOS screen-recording GUI and workflow automation agent built on the Screenpipe CLI. Best for macOS power users who want a free GUI over the Screenpipe CLI, Developers comfortable with git clone, npm install and managing an OpenRouter key, Privacy-first users who need screen history to stay on-device. Free to use.

Compared withvs Locus Robotics vs Truleo vs Presto Voice

What people actually say about ghostwork — is it worth it?

We scanned public community sources for ghostwork on Jul 24, 2026 and could not establish that the discussion we found is about this tool rather than something else sharing its name. Our own analysis of that scan says the posts were off-subject. Rather than publish a sentiment score built on the wrong subject, we publish nothing here and re-run the scan.

Viability Score

69/100

Monitor

How well maintained and how widely used is ghostwork? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this

Recent activity

not measured

Traction

100

Site health

User sentiment

What the vendor publishes

Last calculated: October 2026

How we score →

Key Features

Timeline view for drilling into every recorded Screenpipe session
Natural language search across your screen history
App usage analytics with per-app breakdowns
Learns workflows by observation — no prompts or scripts
Four memory layers: working, episodic, semantic, procedural
L1 working memory polls app, URL, and OCR every 10 seconds
L2 episodic memory ingests clicks, keystrokes and app switches every 2 minutes
L3 semantic rules promoted nightly, e.g. WHEN on LinkedIn search → DO export to CRM
L4 procedural skills compiled with step-by-step DOM/AX locators
Nightly NREM/REM/GC consolidation: promote, compile, decay, dedup, 90-day prune
Autonomy tiers: supervised by default, autonomous after 5 accepts / <2 rejections
AX-first native control via macOS Accessibility API (~95% accuracy on native apps)
Claude vision fallback for browsers and AX-empty apps
Zero-token deterministic replay of compiled skills
Multi-step rollback fires Cmd+Z in reverse order on failure

About ghostwork

FreeIntermediateNo APIDesktop

Ghostwork is a free, open-source Electron app for macOS 12+ that wraps the free Screenpipe CLI in a full GUI, then adds a workflow-automation agent on top. If you have been eyeing Screenpipe's official desktop app since it moved to $25/month, Ghostwork reproduces the same timeline view, natural-language screen-history search, and app usage analytics for $0 — and goes further by automating the repeated multi-step flows it finds. It is aimed at macOS power users who repeat the same clicks across LinkedIn, a CRM, and an inbox every day. The learning model runs on four memory layers. L1 Working holds current screen context (app, URL, OCR) polled every 10 seconds. L2 Episodic ingests raw clicks, keystrokes, navigations, and app switches every 2 minutes from Screenpipe's input stream. L3 Semantic promotes those raw events into rules such as "WHEN on LinkedIn search → DO export to CRM." L4 Procedural compiles stable rules into executable skills with step-by-step DOM/AX locators. A nightly three-phase consolidation (NREM, REM, GC) handles promotion, compilation, decay, dedup, and a 90-day prune, while a living behaviour.md profile is injected into every trigger decision. Autonomy is earned rather than assumed. New rules start Supervised: they execute immediately, show a HUD notification, and stay undoable with Cmd+Z. A rule graduates to Autonomous after 5 accepts and fewer than 2 rejections in the last 10 runs; anything externally visible — sending an email, submitting a form, posting — always stops for one-tap approval regardless of tier. Execution is AX-first through the macOS Accessibility API with a Claude vision fallback for browsers and AX-empty apps; compiled skill replay is zero-token and deterministic, and a failed multi-step sequence fires Cmd+Z in reverse order. Against alternatives, the choice is simple: the paid Screenpipe desktop app is the maintained, vendor-supported timeline tool, while Ghostwork is the only free option that later automates what it

Behind the Verdict

The pitch is narrow and honest: if you already run the free Screenpipe CLI and want a GUI over your screen history, Ghostwork gives you the timeline, natural-language search, and analytics that Screenpipe's official app charges $25/month for. That alone is a fair reason to clone the repo. The bigger bet is automation. Ghostwork watches clicks and navigations, promotes repeated patterns into rules overnight, then compiles stable ones into replayable skills. We'd reach for it when the same LinkedIn-search-to-CRM export or form-filling loop eats an hour a day. Where it bites: execution leans on macOS Accessibility APIs with a Claude vision fallback, so browser-heavy and AX-empty apps are the fragile path. The repo claims roughly 95% accuracy on native apps; assume less on the web. The supervised tiering is the right design. New rules fire once with a Cmd+Z escape hatch, and externally visible actions always wait for one-tap approval, so the blast radius of a bad rule stays small. Rollback only covers reversible steps, though — don't point it at anything you can't take back. Pick it if you're a Mac power user who lives in a terminal, wants screen history to stay local, and is happy filing GitHub issues instead of support tickets. Pass if you need Windows or Linux, cloud sync, shared team memory, or a vendor on the phone. Versus alternatives, the closest is Screenpipe's own $25/month desktop app: the maintained, supported path with no automation. Rewind and similar screen-history tools record but don't act. Ghostwork sits in a small class of record-and-automate agents, and it's free because it's alpha. Our advice: install it on one Mac, let it observe for a week, and only promote rules you'd be comfortable undoing. Treat the automation as a bonus on top of a free

Researching ghostwork? Get your full AI stack in 60 seconds.

Free, no signup — tell us your goal and get tools matched to your budget & existing stack.

Real-world workflow fit

Concrete scenarios for the personas ghostwork actually fits — and what changes day-one when you adopt it.

Solo recruiter running LinkedIn sourcing all day

Ghostwork records each LinkedIn search plus the CRM export that follows, the nightly NREM/REM pass promotes it to a semantic rule ('WHEN on LinkedIn search → DO export to CRM'), and REM compiles it into a procedural skill with DOM/AX locators.

Outcome: The rule runs Supervised first — HUD notification, Cmd+Z available — and after 5 accepts with fewer than 2 rejections it graduates to Autonomous and logs each run to the Activity feed.

Privacy-conscious developer

They run the free Screenpipe CLI locally with Ghostwork's GUI layer on top, use natural-language search over the recorded timeline, and keep screen history on-device with no cloud account.

Outcome: They get the timeline view, natural-language search and app usage analytics that Screenpipe's $25/month desktop app charges for, at $0 and self-hosted.

Ops generalist tired of form filling

A multi-step browser submission is learned from observation and replayed by the compiled skill at zero tokens; when a step fails mid-sequence, Ghostwork fires Cmd+Z for each completed reversible step in reverse order before surfacing the error, and the externally visible submit still waits for one-tap approval.

Outcome: Repetitive form filling moves off their plate without a scripting project, and nothing irreversible goes out without their approval.

Use Cases

Search your screen history in natural language to find what you were doing last Tuesday
See per-app usage analytics to understand where your Mac time actually goes
Let the agent learn a repeated LinkedIn-search-to-CRM export and run it for you
Automate inbox triage and form filling from observed patterns rather than scripts
Replace the $25/month Screenpipe desktop app with a free open-source GUI over the same CLI
Run screen recording and workflow learning locally, with no cloud account

Models Under the Hood

Claude vision fallback

as of 2026-10-11

Limitations

Ghostwork is 0.1.x-alpha: 20 commits, 106 stars and 12 open issues at the repo scrape, so expect gaps and bugs.
Setup is manual — git clone, npm install, Node.js 20+, the Screenpipe CLI and an OpenRouter API key.
The app is free, but the Claude vision fallback for browsers and AX-empty apps routes through a paid LLM provider on your key, so heavy browser automation has a real per-call cost.
Rollback only covers undoable actions, so irreversible steps are unsupported. macOS 12+ only — no Windows or Linux build.
No cloud sync, shared memory or admin controls for teams.
Solo-maintainer, community-driven project with no support contract.

as of 2026-09-27

Verification history

We have re-verified ghostwork 8 times since Jul 3, 2026. Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.

Oct 10, 2026 — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Sep 24, 2026 — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Sep 1, 2026 — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Aug 19, 2026 — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Aug 3, 2026 — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Jul 24, 2026 — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it

Showing the 6 most recent of 8 verification passes.

Free to cite with attribution — this page re-verifies continuously.

12-month cost

Project the real annual outlay, including the implied monthly cost when only an annual tier is published.

Plan

Annual total

Free

Over 12 months

Effective monthly

—

Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.

Plans compared

For each published ghostwork tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.

Free (open source, GPL-3.0)

Hidden costs & gotchas

What the public pricing page doesn't put in bold. Captured from pricing-page footnotes, contract terms, and recurring complaints.

The app is free, but browser and AX-empty workflows route through the Claude vision fallback, so each fallback call bills against your own OpenRouter or Anthropic key.
You must supply and pay for an OpenRouter API key for the nightly LLM consolidation that promotes episodic events into semantic rules and compiled skills.
Running the free Screenpipe CLI underneath means Screenpipe keeps recording to local disk, and the GC phase prunes only raw events older than 90 days — storage grows until then.
Self-hosting means your time is the maintenance budget: Node.js 20+, Screenpipe, and repo updates are all unpaid work you do yourself.

Where the pricing makes sense

The company stage and team size where ghostwork's pricing actually pencils out — and where peers do it cheaper.

Setup time & first value

How long it actually takes to get something useful out of ghostwork — broken out by persona, not the marketing-page minute.

Switching to or from ghostwork

How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.

Migrating in

→From Screenpipe desktop app ($25/mo): point Ghostwork at the free Screenpipe CLI and use its GUI layer for the same timeline, search and analytics.
→From manual browser scripting: stop maintaining selectors — let the L2 episodic stream record the flow and the nightly consolidation compile it.
→From a paid RPA seat: replace scheduled click sequences with observation-learned skills, accepting that Ghostwork is macOS-only and alpha.

Migrating out

↗To Screenpipe desktop ($25/mo): if you want the recorded timeline, search and analytics with a maintained vendor app and no source build.
↗To a commercial RPA or desktop agent: if you need support SLAs, Windows/Linux coverage, or team admin controls.
↗To doing it by hand: if you decide the alpha code and OpenRouter key aren't worth the maintenance.

Integrations

ScreenpipeOpenRouterAnthropicLinkedIn

Resources & Guides

Tutorials & Learning

YouTube returned 6 videos for “ghostwork”, and we withheld 6: 6 could not be judged, because “ghostwork” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about ghostwork.

Official links

Official Website Changelog

Tools that pair well with ghostwork

Common stack mates teams adopt alongside ghostwork, with the specific reason each pairing earns its keep.

Chrome DevTools MCP

Open-source MCP server that gives coding agents live Chrome DevTools access for debugging, browser automation, and performance traces.

Open Interpreter

Open Interpreter is an open-source terminal coding agent that turns plain-English requests into real file edits and shell commands.

Stagehand

Open-source SDK for building browser agents with AI primitives (Act, Extract, Observe, Agent) plus Playwright-style controls in TypeScript, Python, and Go.

Featured Head-to-Head Comparisons

Ghostwork vs Locus Robotics

Choose Locus Robotics if you run a warehouse needing physical automation to boost picking productivity; it's enterprise-grade with RaaS. Choose ghostwork if you're a macOS developer or power user wanting a free, privacy-focused tool that automates repetitive computer tasks by learning your workflow. They serve completely different domains—no direct competition.

Ghostwork vs Truleo

Choose Truleo if you need a law enforcement‑specific intelligence platform connecting RMS, CAD, jail calls, and BWC. Choose Ghostwork if you want a free, privacy‑first macOS automation agent that learns your desktop workflows. They serve completely different users and use cases.

Ghostwork vs Presto Voice

These tools serve completely different markets and needs—Presto Voice is an enterprise-grade drive-thru AI for QSR chains looking to boost revenue, while ghostwork is a free, open-source desktop automation agent for macOS developers. Choose based on your domain: restaurants or personal productivity. They are not direct competitors.

Alternatives to ghostwork

View all

Frequently Asked Questions

Best-of guides

Best AI Tools for Productivity Best AI Meeting Assistants & Note Takers Best AI Tools for Project Management Best AI Note-Taking & Knowledge Management Tools

Topics

Automation Agent API Data Analysis Open Source

Used ghostwork? Help shape our editorial sentiment research.