Pricing
Understand the costs of running AI coding agents on Superconductor.
When using AI coding agents, there are two things to pay for: tokens and compute.
Tokens
Tokens are the currency of large language models. Every time an agent reads code, thinks about a problem, or generates a response, it consumes tokens.
You can bring your own tokens — connect your own API keys, or your existing Claude, ChatGPT, or Grok Build plan. See Workspace billing and credits for details.
Compute
Compute is the cloud infrastructure where your agents run. Each implementation spins up an isolated environment with CPU, memory, and storage.
Compute is measured in sandbox hours. One sandbox hour is one hour of a 1 vCPU / 2 GB RAM instance — a larger environment uses proportionally more. Usage is billed per second, and instances pause automatically when idle, so you only pay while an agent is actively working or you're using a live preview.
Plans
Superconductor has Free, Pro, and Enterprise plans. Every plan includes the full multiplayer workspace; the plans differ in their monthly sandbox hours, environment sizes, team members, and concurrent agents.
See the pricing page for current plans, sandbox-hour allotments, and rates.