always know what you pay.

one subscription: a cloud machine that keeps your work and wakes on a message, every model at the provider's price, and agents that keep working while you're away. usage shows live — never a surprise bill.

max

$200/mo

a fleet: up to three large machines, five agents in parallel.

  • $200 of usage a month
  • 12 slots · all sizes, up to 3 large
  • 60 GB shared storage
  • 20 live urls

a medium and a small warm all month are $52.56 — about $147 stays for models.

go max →
team

$500/mo

one pool for the whole team, unlimited members.

  • $500 of usage a month, pooled
  • 60 slots · up to 15 large
  • 100 GB shared storage
  • 50 live urls

optional per-member caps; every member draws from the same pool.

create a team →

free — the hub from your own tools: every model in the cli, your editor or any OpenAI-compatible client, pay as you go from prepaid credits at the provider's price. top-ups carry a 3% fee and the key pauses at $0. open the hub →

new here? try pro for a week for $12, then $50/month. see the trial →

provider price models every model in the catalog, on every plan, at the provider's price with no markup.
$0.04/vCPU-hr · $0.012/GB-hr machine vCPU while working, memory while warm, nothing while asleep.
same prices past your allowance prepaid credits pick up automatically — no overage rate, no surprise.

compare plans

feature
free $0
pro $50/mo
max $200/mo
team $500/mo
models (hub)
included models
every model — pay as you go from credits
every model in the catalog
every model in the catalog
every model in the catalog
usage allowance
none — pay as you go from prepaid credits
$50 a month
$200 a month
$500 a month — one shared pool
model pricing
provider price
provider price
provider price
provider price
credits
3% top-up fee
3% top-up fee
3% top-up fee
3% top-up fee
cloud
machine slots (1 slot = 1 vCPU)
—
3 — small, medium
12 — all sizes, up to 3 large
60 — all sizes, up to 15 large
developer api
—
✓ 180 req/min
✓ 300 req/min
✓ 600 req/min, pooled
octos (agents with their own memory)
—
10
30
30
concurrent sessions per machine
—
10
20
20
shared storage
—
20 GB
60 GB
100 GB
image attachments (it sees your screenshots)
—
✓
✓
✓
voice in and out (speak, and be spoken to)
—
✓ $0.010/min in · $0.025/min out
✓ $0.010/min in · $0.025/min out
✓ $0.010/min in · $0.025/min out
media generation (images, video, audio)
—
✓ from $0.004 / image · $0.04 / second video · $0.002 / second audio
✓ from $0.004 / image · $0.04 / second video · $0.002 / second audio
✓ from $0.004 / image · $0.04 / second video · $0.002 / second audio
Web search (your agent looks things up live)
—
✓ $5 per 1,000 searches
✓ $5 per 1,000 searches
✓ $5 per 1,000 searches
Browser starts (your agent looks things up live)
—
✓ $0.02 / browser
✓ $0.02 / browser
✓ $0.02 / browser
Browser time (your agent looks things up live)
—
✓ $0.015 / minute
✓ $0.015 / minute
✓ $0.015 / minute
Own-key requests (your agent looks things up live)
—
✓ $0.05 per 1,000 requests
✓ $0.05 per 1,000 requests
✓ $0.05 per 1,000 requests
custom Docker images
—
up to 5
up to 20
up to 50
live urls for your apps
—
5
20
50
url visibility
—
private by default, public when you choose
private by default, public when you choose
private by default, public when you choose
suspend after idle
—
1 hour default — up to always on
1 hour default — up to always on
1 hour default — up to always on
machine retention
—
kept — never reclaimed
kept — never reclaimed
kept — never reclaimed
reach it from Telegram, Slack, WhatsApp
—
✓
✓
✓
artifacts (files that outlive the machine)
—
✓
✓
✓
routines (scheduled agent runs)
—
✓
✓
✓
cold archive (data kept, restores in minutes)
—
after 3 days suspended · extra disk $0.04/GB-mo
after 3 days suspended · extra disk $0.04/GB-mo
after 3 days suspended · extra disk $0.04/GB-mo
team
members
—
—
—
unlimited — join by invite link
parallel model calls
1 · 2 after any top-up · up to 5
5
10
20 (pooled)
per-member limits
—
—
—
optional owner-set caps

credits

prepaid balance that picks up automatically once your plan allowance is spent, so a job never dies mid-work. same prices as your plan allowance: models at the provider's price, machine time, storage and media at their listed rates. a top-up adds 3% on top: add $100, pay $103.

$5+minimum top-up
3%added when you top up
∞credits never expire

model prices come from the octomind version we run — metering always matches the published table. cancel any time; your data stays safe for 30 days — nothing is deleted during the grace period.

the hub is open to every account — use it from your own tools and pay as you go from credits. paid plans add the cloud: a machine that keeps your work and wakes on a message, routines, connectors and live urls.