Back to Home

Pricing

Open Company runs on your own machine, so there are no seats to count and nothing metered on our side. What follows is who bills you for what.

Beta

Open Company is in beta and pricing has not been set. Nothing on this page is a price or a commitment to one — it describes where the costs of running it come from today.

The software
Open sourceMIT licensed

The whole thing, not a trial of it.

Get started
  • Every node and every provider — nothing is held back for a paid tier
  • No subscription and no usage limits
  • No seats to count: run it for one person or for everyone
  • Self-hosted, so your documents and credentials stay on your machine
  • Fork it, change it, ship it — MIT terms, including commercial use

Open source under the MIT licence.

Your models
Your providerbilled directly to you

The only cost that scales with how much you use it.

Model setup
  • Bring your own API keys across 13 providers
  • Or run locally: Ollama and LM Studio are first-class, with no API to pay for
  • Local context length is detected from your own running server
  • Agent runs report USD spend from provider-reported usage, in the credentials panel
  • Set your own rates in pricing.json for models not priced by default

Paid to your provider, never to us.

Your infrastructure
Your machinewherever you install it

Runs where you put it. No tenancy, no egress.

Install guide
  • Node.js 22 or newer and Python 3.12
  • One command: npm install -g @zeenie-ai/opencompany
  • Your laptop, your server, or your own cloud account
  • No vendor sits between you and your business data
  • Community support on Discord and GitHub issues

Often the machine you already have.

Who bills you, and for what

ItemBilled by
Open CompanyNobody today — open source under MIT
LLM tokensYour model provider, at their rates
Local modelsNobody — Ollama and LM Studio run on your hardware
The machineYour host, or nobody if you already run one
SupportNobody — Discord and GitHub issues

Pricing questions

It is in beta, and pricing has not been set — so rather than quote a number that might change, this page sets out who bills you for what. Today the software is open source under the MIT licence and the repository states plainly that there is no subscription and no usage limit.

No. Because you run it yourself there is nothing on our side to meter — no seat count, no per-run fee, and no cap on how many workflows you build or how often they fire. That is the main way the cost model differs from hosted automation tools, which bill per task or per user.

Whatever your model provider charges for the tokens you consume, plus the machine you run it on. Point it at Ollama or LM Studio and the models run on hardware you already own, so there is no API bill at all.

Yes. The MIT licence permits commercial use, modification and redistribution, with no revenue threshold and no separate commercial licence to buy.

Not today. Self-hosting is the intended deployment, which is what keeps your documents, customer records and credentials on infrastructure you control.

Partly. Agent runs connected to memory calculate USD cost from provider-reported usage, where the provider reports it, and that spend is visible in the API credentials panel. It is a spend indicator rather than a complete audit log of every request.

Need something else?

If you are weighing this up for a team and want to talk through how it would be deployed, get in touch. For how it compares with per-seat and per-task tools, see the comparison on the home page.