AI assistants and search decision
Win.sh
A capable developer can build a useful self-hosted subset (connectors, runner, LLM orchestration, memory, approvals) within a multi-week effort, but recreating Win.sh's polished integrations, reliability, and hosted product experience is larger operational work better left to the vendor.
Visit website↗$50/mo
$600/yr
Read off the official pricing page.
$100one-off50 h to build
$100/mo6 h/mo upkeep
On cash alone, building overtakes the subscription at 3 seats.
Open-source builds that already do this
Every project below is open source and already does this job today. Fork one, self-host it, or take the parts you need - the build prompt further down assumes an empty file, and this is the shortcut past that. Licences differ; check the one on each card before you ship. All Win.sh alternatives, with the arithmetic →
What a replacement has to do
- Poll connected services for signals → load company state and memory → run agent loop (LLM calls) to propose or execute actions → record receipts, decisions and updated memory → surface approvals for risky moves
What it still won’t have
- Polished multi-service integrations and maintained connector catalogue
- Enterprise SLAs, high-availability hosting, and full compliance controls
- Large-scale training data or proprietary models and brand trust
- Polished UX around approval flows and reporting dashboards
What remains hard
- Product polish and ongoing maintenance
First-year cost
Keep paying
Paying is—cheaper in year one.
On cash alone, building overtakes the subscription at 3 seats.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a minimal self-hosted 'autonomous company' harness in Node.js (Express) with a small React admin UI. Core features in scope: 1) connectors: implement webhook and polling connectors for Stripe and a generic analytics API; 2) scheduled runner that executes a daily 'run' evaluating signals; 3) LLM orchestration that formats prompts and calls OpenAI/compatible APIs and parses structured actions; 4) append-only JSONL company memory and rules store on disk or Postgres; 5) approval UI showing proposed actions with approve/reject and an audit trail; 6) budget guardrail tracking spend per run and pausing non-critical actions. Out of scope: polished marketplace of integrations, multi-tenant billing, enterprise SLAs, or proprietary model training. Require error handling, retries for external calls, unit tests for runner and connector logic, and basic integration tests for end-to-end runs.
How we checked
How the score was reached
- Partly verdict base52
- An open-source build was found+5
- 5 cited sources+3
- Price verified on pricing page+3
- 3/3 assessment runs agreed+4
- Evidence score67
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time - so the same evidence always produces the same number.
How scoring works →Cited sources · 5
Every page the run actually retrieved.
- official productWin.sh homepage
- official pricingWin.sh pricing
- official docsWin.sh docs
- open sourceCherryHQ/cherry-studio
- open sourcearc53/DocsGPT
Integrity checks
What held up, and what did not.
