Coroid bills two separate things:
- Capacity — agent slots, charged per slot per month.
- Model usage — the tokens your work consumes, either prepaid as credits or billed directly by your own provider.
Keeping them separate is deliberate. Capacity is predictable and budgetable; model usage varies with what you actually run.
Current prices are on the pricing page.
Free
For evaluating the product end to end on real but lightweight work.
Includes one project with 1 GB of storage, one low-priority agent slot, 2 agent-hours a day (UTC) within 20 a month, a ceiling on how long a single task may run, and no parallel subagents. Models are Coroid-routed at no model cost. The organization owner is the only member; inviting teammates needs Professional.
It is a genuine evaluation tier — the whole governed workflow runs, including specifications, plans, gates and review. What it limits is throughput, not capability.
Professional
For teams running standard workflows.
A flat monthly fee per organization including one standard agent slot, with additional slots available per month. Unlimited projects and members, with 10 GB of project storage. Bring your own provider keys, or prepay model credits. System and custom AI profiles.
Members are unlimited because seats are the wrong unit for this product — ten people sharing one slot still get one task at a time. You pay for work done, not for people who might request it.
Enterprise
For larger organizations with contracted capacity and deployment requirements.
Contracted slots with a minimum commitment (currently five), 50 GB of project storage, priority processing, custom AI profiles with per-agent model choice, SSO/SAML with SCIM provisioning, audit log export, a dedicated support SLA, and an on-premises option.
Model pricing
Coroid applies no markup to provider token rates. Under BYOK you pay your provider directly at your own rates; with prepaid credits you pay provider rates through Coroid.
See Model credits.
Security on every plan
Security runs on every plan, Free included. Detection (dependency inventories and advisory matching, secret and code scans, and GitHub alert imports) uses no AI model, so it costs neither credits nor agent hours; it takes one of your agent slots while it runs. Work that uses a model, such as fix pull requests, the security audit and the independent check of security issues, spends credits and agent hours like any other work. The security audit is off on Free until you switch it on.
Choosing a plan
Start free and run real work. The question that decides whether you need Professional is not "is this good enough?" but "is my queue backing up?" — that is a capacity question, and capacity is what you are buying.
Enterprise is for the deployment and identity controls, not for capacity — additional slots are available on Professional. Since it is in private preview, talk to us before planning around it.