Pricing

Pay for the platform and compute. Your models stay on your accounts

Connect your own Anthropic, OpenAI, Gemini, Vertex AI or self-hosted accounts. kosem never marks up tokens and charges nothing for agent model calls.

Hobby

$0 a month, plus usage

$10 of credits to start

For one engineer running coding agents on their own.

Start free
  • $10 of starter credits for 60 days (~43 hours of Standard environments)
  • Add a card anytime to pay as you go, with no subscription
  • 1 engineer · 3 projects · 4 concurrent vCPUs (2 Standard environments)
  • Agents, LLM Gateway, dynamic credentials and private sites
  • 2 model providers and 2 vault connections
  • Agent model calls included ($0) · 7-day request log

Pro

$15 a month, plus usage

For solo power users

For one engineer running larger environments and custom model endpoints daily.

Contact us
  • Everything in Hobby, and:
  • Environments up to 8 vCPUs, 52 GiB memory and 500 GiB disk
  • 16 concurrent vCPUs, 64 GiB memory and 15 projects
  • 10 model providers, including custom and self-hosted endpoints
  • 10 vault connections and 10 prepared images (100 GiB)
  • 30-day request log and 50 sites
  • Email support

Enterprise

Custom annual contract

Volume rates and SLA

For companies standardizing coding agents across teams on custom terms.

Contact sales
  • Everything in Team, and:
  • Volume discounts on cloud compute and storage
  • Optional annual credit commitment
  • Custom quotas and dedicated capacity limits
  • Invoices and purchase orders
  • Support SLA and a named contact

Prices in US dollars, before tax. kosem's team sets up Pro and Team for now, and helps size any deployment: write to kosem's team. Compare plans

Usage

Cloud environments. Billed per second while running

Compute is $0.0504 per vCPU-hour and $0.0162 per GiB-hour while running. Disk is $0.10 per GiB-month until deleted. Hobby, Pro and Team pay the same list rates from credits.

  • Small1 vCPU · 4 GiB$0.115 an hour
  • Standard2 vCPUs · 8 GiB$0.230 an hour
  • Large4 vCPUs · 16 GiB$0.461 an hour
  • Extra large8 vCPUs · 32 GiB$0.922 an hour

Agent model calls are free on every plan. Direct LLM Gateway API calls include 50,000 (Hobby), 250,000 (Pro) or 1,000,000 (Team) requests a month, then $0.10 per 1,000. Agent web searches include 100 (Hobby), 500 (Pro) or 1,000 (Team) a month, then $0.012 each.

  • Solo engineer on Hobby~$17 / month

    3 hours a workday on Standard · 20 GiB disk

    $0 platform + ~$17 compute and disk
  • Solo power user on Pro~$34 / month

    3 hours a workday on Standard · 40 GiB disk

    $15 Pro + ~$19 compute and disk
  • 10-builder team on Team~$470 / month

    10 builders × 3 hours a workday on Standard · 5 free viewers

    $300 builders + ~$170 compute and disk · $0 for viewers

Compare

Compare plans. Start solo and add governance when a team joins

The limits and controls that distinguish Hobby, Pro, Team and Enterprise. Full technical quotas are in the limits reference.

FeatureHobby$0 a month, plus usageStart freePro$15 a month, plus usageContact usTeam$30 a builder a month, plus usageContact usEnterpriseCustom annual contractContact sales
Seats and scaleOne engineer on Hobby and Pro; unlimited builders with free viewers on Team.
BuildersAnyone with permission to run agents, environments or changes11$30 / month eachCustom
Viewers, billing and audit rolesRead-only viewers, site viewers, billing admins and audit log viewersNot includedNot includedFreeFree
Projects31550raised on requestCustom
Concurrent vCPUsMemory scales at 4 GiB per vCPU (16 / 64 / 200 GiB)42 Standard environments1650raised on requestCustom
Largest environment2 vCPUs · 8 GiB · 50 GiB disk8 vCPUs · 52 GiB · 500 GiB disk8 vCPUs · 52 GiB · 500 GiB disk8 vCPUs · 52 GiB · 500 GiB disk
Total environment disk100 GiB500 GiB2,000 GiBraised on requestCustom
Platform and integrationsConnect your own model providers, cloud accounts and private images with zero token markup.
Agent model callsModel tokens stay on your own provider billsUnlimited ($0)Unlimited ($0)Unlimited ($0)Unlimited ($0)
Direct LLM Gateway API callsFor your own apps and scripts calling the gateway directly50,000 / monththen $0.10 per 1,000250,000 / monththen $0.10 per 1,0001,000,000 / monththen $0.10 per 1,000Volume rates
Model providersAnthropic, OpenAI, Google Gemini and Vertex AI21025org-shared across projectsCustom
Custom and self-hosted model endpointsOpenAI-compatible and Anthropic-compatible serversNot includedIncludedIncludedIncluded
Vault connections (GitHub, AWS, GCP)Short-lived dynamic credentials scoped per environment21025org-shared across projectsCustom
Prepared imagesPrivate images cached for fast environment starts110 GiB storage10100 GiB storage25500 GiB storageCustom
SitesStatic sites published by agents, private or public (5 / 20 / 50 recent versions kept)530-day inactivity retention50200raised on requestCustom
Agent file storageTotal files attached to agents across the organization1 GiB25 GiB250 GiBraised on requestCustom
Governance and auditShared collaboration, org-wide guardrails, data retention and immutable audit trails.
Shared agents and org-shared resourcesNot includedNot includedIncludedIncluded
Groups and fine-grained RBACNot includedNot includedIncludedIncluded
Organization policies and custom retentionAllowed tools, max sizes, public site controls and org or project retention limitsNot includedNot includedIncludedIncluded
Agent and file retentionHow long inactive agents, transcripts and attached files stay; daily usage totals stay90 days365 daysUnlimitedshortenable by policyCustom
Stopped environment retentionMaximum time stopped environments and their disks stay before deletionUp to 7 daysUp to 30 daysUnlimitedshortenable by policyCustom
LLM Gateway request logPer-request log without prompts or answers; daily sums stay7 days30 days90 daysshortenable by policy90+ days
Audit logs (Admin Activity and Data Access)Recorded on every plan for 400 days; Team unlocks viewing and Data Access controlsNot includedNot included400 days400+ days
Billing and supportTransparent list rates on Hobby, Pro and Team; custom contracts on Enterprise.
Cloud compute and storage ratesList pricesList pricesList pricesVolume rates
Monthly usage limits and auto-reloadIncludedIncludedIncludedIncluded
Payment and supportCard · DocsCard · EmailCard · EmailInvoice / PO · SLA and named contact

Looking for per-resource limits on secrets, registries, service accounts and rate quotas? See Limits and quotas in the docs.

Questions

How billing works. In a sentence or two each

Does kosem charge for model tokens?

No. Agents call models through your own Anthropic, OpenAI, Google Gemini, Vertex AI or self-hosted accounts, on your contracts, and your providers bill you directly. kosem never marks up tokens and never charges per request for agent model calls.

Who counts as a builder seat on Team?

A builder is any member or guest with a role that starts or steers agents, manages environments or changes resources. Read-only viewers, site viewers, billing admins, member admins and audit log viewers are free. Seats are counted once a month at renewal, with no mid-month charges.

What happens after the $10 of starter credits run out on Hobby?

Add a card in Billing to keep paying only for the cloud environments and storage you use, with no monthly subscription. Starter credits last 60 days and cover about 43 hours of Standard environments.

What do stopped environments cost?

Only their disk, at $0.10 a GiB-month until deleted: a 20 GiB disk costs about $2 a month. vCPUs and memory are billed per second only while running, and an agent's environment stops 5 minutes after its turn.

Can an organization cap what it spends?

Yes. On Hobby, Pro and Team, set a monthly usage limit in Billing › Overview to stop paid usage at your cap, and set a monthly ceiling on credit auto-reload. The LLM Gateway also enforces monthly spend limits on your model provider bills.

Does kosem rate-limit how fast agents call models?

Your providers' own rate limits apply, and their 429s reach your agents directly. kosem doesn't throttle model speed by plan.

Start with $10 of credits

Start free