OmniLink
Get OmniSim
OmniLink
Free agents run while you’re connected. Paid agents run when you’re not. Every plan — including the free one — gives you unlimited agents, the full platform, and your own model keys. You pay your model provider directly; we never take a cut of your tokens.
Build as many agents as you like and run them while you’re connected.
Three Pro agents that keep working after you log off — each commanding a team of up to 4.
Fifteen Pro agents running unattended around the clock, each commanding a team of up to 10.
Sixty Pro agents with teams of 25, and a direct vote on what gets built next.
Prices are in US dollars and exclude any tax your jurisdiction adds at checkout. Paid plans are billed monthly in advance and renew automatically on the same day each month until you cancel — there is no annual option and no minimum term. Payment is handled by Stripe; we never see your full card number.
You can cancel at any time from the dashboard. Cancelling stops the next renewal and leaves you on the paid plan until the end of the period you have already paid for, after which you drop to Free. We do not pro-rate or refund the remainder of a period you have started, except where the law where you live requires it — if a charge looks wrong, write to support@omnilinktechnology.com within 30 days and we will sort it out.
Two things the subscription is not. It is not metered: there is no usage-based charge, no overage and no credit balance on any tier. And it does not include model inference — you connect your own provider key and that provider bills you directly, at their prices, with no markup from us.
Two things. First, whether we will wake your agent on a schedule while you are not there — that is the work our infrastructure actually does. Second, how large a team a Pro agent may command through delegation: 4 on Builder, 10 on Master, 25 on Lord, where one order fans out to the whole team in parallel. Everything else — building agents, wiring them to OmniSim, talking to them, the Secretary, the live cost breakdown — is uncapped on the free plan, because it costs us almost nothing.
Be precise about what persistence means here. The agent runs on
your machine through the OmniLink connector, which is why your
tools, files and robots never leave your network, and why a robot
bridge on 127.0.0.1 is reachable at all. A scheduled agent
survives closing this tab and logging out. It does not survive the
machine being switched off, or the connector not running — there
is nothing on your side to execute the work. Its conversation memory
lives with us, but that is a rolling window of recent turns, not an
archive: old turns drop out, and there is no long-term semantic recall
behind it.
OmniLink is bring-your-own-key, and you should hear that here rather than discover it after signing up. A Google AI Studio key takes about a minute and needs no card; OpenRouter connects in one click with free open models; a local Ollama needs no account at all. Already pay for OpenAI, Anthropic or xAI? Paste that instead.
We never mark up a token, at any volume, on any plan. Your provider bills you directly. The subscription is the only money that reaches us.
OmniSim is an Apache-2.0 simulator you run yourself. It is not metered, nothing on this page limits it, and we do not charge for simulated robot time. What an agent claims it did gets checked against the simulated physics rather than taken on its word.
The usual way to price an agent platform is per minute of agent time. Vapi and ElevenLabs both meter it that way — Vapi even when you bring your own model key. That bills you for the model talking, which is already your token spend. We charge a flat monthly fee for the one thing we actually run: waking your agent while you are away. Your key, your quota, your bill.
Any agent configured to run without you present: a standing order on a schedule, an agent woken by a webhook, or one listening on a connected channel. An agent you talk to in the browser does not count against the limit, however long the conversation runs. (These were called “persistent agents” until July 2026 — same thing, same enforcement.)
The Secretary is the platform watching an agent’s connection around the clock. When the link drops, the agent messages you — in its own voice; when the link returns you get an away-brief naming what it could not do and what needs your confirmation; once a day it can post a digest of status, activity, incidents and cost. It runs on every plan because it is the clearest demonstration of what this platform is — the paid tiers gate scale, not honesty.
No plan limit — every tier, including Free, has unlimited agent profiles, and upgrading never changes that. One honest footnote: an anti-abuse ceiling of 1,000 profiles per account exists so a script can’t flood the platform with junk rows. It is the same on every plan (so it is not something you can pay your way past), no human use comes anywhere near it, and if you genuinely need more, contact us and we’ll raise it.
Each plan keeps memory and history for a window: 30 days on Free, 90 days on Agent Builder, a year on Agent Master, forever on Agent Lord. What ages out past the window: old messages, idle conversations, the memory of agents you haven’t talked to inside the window, and run history. What never expires, on any plan: your agents themselves — profiles, settings and schedules are configuration, not history. An agent you talk to regularly keeps its memory alive on every tier.
One honest caveat about today. The sweep that implements those windows is currently running in report-only mode: it counts what would age out and deletes nothing. So in practice history is still being kept, on every tier, and the windows above describe what we will apply rather than what has already happened. We will announce a date before that changes, so nobody loses history to a policy they were not told was starting.
And the direction that catches people: the window follows your current plan, not the plan you were on when the history was made. Downgrading shortens it, and once enforcement is live, history older than the shorter window goes. Export or copy anything you want to keep before you drop a tier.
You are told which limit you hit and when it resets. Nothing is silently dropped and nothing is billed as overage — there is no usage-based charge on any tier, so going over costs you a wait, never money.
No. Agent Builder, Master and Lord are the same plans at the same prices you signed up at. The old per-plan cap on how many agents you could create is gone on every tier, so you simply gain the uncapped profiles — and where per-minute limits changed, they only went up. Nothing to migrate; your price is untouched.
Yes, from the dashboard, at any time. You drop to the free plan at the end of the period you have paid for, and your agents stay where they are.