Every seat includes the full platform and monthly credits for Agent Computer runtime. Bring your own API key or connect ChatGPT for model access. Managed models remain available and use credits based on token usage.
Kortix credits are primarily for Agent Computer runtime. Use your own API key or ChatGPT subscription to keep model billing with your provider. Optional managed models use Team credits based on token usage.
Compute is billed by the second. The default Agent Computer has 2 vCPU, 4 GiB RAM, and 20 GiB storage. It uses about 20 credits ($0.20) per hour. Auto-stop pauses compute charges, so stopped time uses 0 credits. Optional managed model usage is token-based and uses the same pooled Team credits.
Free includes 200 credits each month for Agent Computer runtime. Those credits do not pay for managed LLM calls.
Bring your own API key or connect ChatGPT. You pay your model provider directly and keep Kortix credits for Agent Computer runtime.
Billed per resource, per second: $0.0000168/vCPU, $0.0000054/GiB RAM, $0.000000036/GiB storage. The default 2 vCPU / 4 GiB / 20 GiB machine runs about $0.20/hour, and auto-stops when idle so you pay $0 the moment it’s not running.
200 credits covers sandbox runtime for early projects and demos.
Kortix-managed models remain available when you need them. Their token-based usage draws from pooled Team credits.
Each Team seat includes 2,500 pooled credits. Used only for compute, that covers about 125 hours on the default Agent Computer.
Free includes 200 credits each month for sandbox compute and 3 projects. Bring your own API key or connect your ChatGPT subscription for premium access. Managed Claude, GPT, and Gemini on Kortix keys are paid.
$40/seat/month includes 2,500 pooled credits per seat, optional managed model access, and seats for the people on your team. Agent Computer runtime and managed model token usage draw from the same pool.
Agent Computer compute is billed per second, per resource — $0.0000168/vCPU, $0.0000054/GiB RAM, $0.000000036/GiB storage — about $0.20/hour for the default 2 vCPU / 4 GiB / 20 GiB machine, and $0 while stopped. Bring your own key or connect ChatGPT to pay your model provider directly. If you choose Kortix-managed models, their input, output, and cached tokens use Team credits at that model’s rate. Free credits remain sandbox-only.
Both. The seat is a flat monthly fee that includes credits. Top up only when Agent Computer runtime or optional managed model usage exhausts the pooled balance.
Everything in Team plus SAML SSO, SCIM directory sync (Okta, Microsoft Entra, JumpCloud), advanced RBAC, audit logs, an SLA and DPA, and Cloud / VPC / on-prem deployment. Talk to us for volume pricing.
$40 per seat, with $25 of usage credits each. Auto-prorated, cancel anytime — or talk to us about Enterprise.