Glubean

Pricing

Local is free forever. Cloud is $12 a seat.

Run tests, contracts, and load on your own infra — no per-run, no per-VU, no surprises. Pay per editor for the cloud workspace; readers are always free. Every new workspace starts with 30 days of Pro — no card.

Free

Bring the whole team in.

$0
  • 2 editors · unlimited viewers
  • Run on your own infra — unlimited, no per-VU
  • Team sync & API design — free (author with a seat)
  • Recent-runs dashboard · 14-day history
  • 200 agent credits / mo (GPT-5.4-mini)
  • Community support

Pro

Popular

Per editor. From one seat, no minimum.

$12/editor/mo
  • 30-day free trial — up to 10 editors, no card
  • Everything in Free, plus:
  • Private team dashboard · 1-year retained history
  • Cross-run intelligence — regression & flaky per assertion
  • Alerts · PR/CI checks · ownership & notifications
  • 1 pack / editor: 1M events + 5 GB — add packs for heavy CI
  • Unlimited editors · unlimited workbench
  • 300 agent credits / editor (pooled) · BYO key
  • Email support

Cloud pricing is in beta — numbers may change. The Pro trial needs no card and downgrades to Free automatically; nothing is deleted. Viewers are always free — you're never charged per person to read a run.
Enterprise (SSO, audit, self-host, SLA) is coming — register interest.

Transparent by design

You can see exactly what you pay for.

Flat per editor

An editor is a seat an admin assigns. Authoring needs a seat — but nobody is auto-charged for clicking edit. Viewers read free forever; switch anyone back to read-only anytime.

Agent metered in credits

1 credit = 1¢, billed by what the model actually uses. Or bring your own key for unlimited.

Cloud capacity is included, not infinite

Each plan includes a monthly evidence allowance, artifact storage, and a retention window — so heavy CI can't surprise either of us. Running on your own infra stays unlimited and free.

Never metered

No per-run execution fee, no per-VU load bill, no charge for read-only viewers.

Compare plans

Everything an editor unlocks — code or no code.

Viewers read for free. Become an editor — write tests, build flows, or design API specs — and the whole workspace opens up.

FeatureFreePro
Workspace
Free trial (no card)—30 days · 10 editors
Editors (anyone who authors)2Per seat · $12/mo
Read-only viewersUnlimitedUnlimited
Private team workspace
Design — APIs (writers, PMs)
API specs3Unlimited
Endpoints250Unlimited
Shared schema components80Unlimited
Writer overlay — prose & examples
Version freeze
Explore — Collections (QA, support)
Requests & flow nodes500Unlimited
Contract-validation in flows
Team-synced collections
Verification — runs & dashboard (devs)
Run test / contract / load on your infraUnlimitedUnlimited
No per-VU · no per-run execution bill
Cloud target dashboardRecent-runs previewFull team dashboard
Cloud history retention14 days1 year
Cross-run intelligence — regression & flaky
Failure repro (request / response)RecentFull + retained
Evidence capacity (events + artifacts)100k + 1 GB1 pack / editor (1M + 5 GB)
Agent
Included credits / mo200300 / editor
ModelsGPT-5.4-miniAll incl. Opus
Bring your own key
Collaboration & operations
Alerts
PR / CI checks
Ownership & notifications
SupportCommunityEmail

Free forever

The whole toolkit, on your own machine.

@glubean/sdk, CLI, and the VS Code Extension

Run tests, contracts, and load on your own infra

Local runs and evidence — nothing leaves your machine

Read-only viewers, on every plan

Load testing without the per-VU cloud bill. You run the load on your own infrastructure; Cloud just stores and analyzes the results. No virtual-user-hour metering, ever.

Evidence capacity

Your CI evidence, in capacity packs.

Each editor seat includes one capacity pack — ~1M evidence events + ~5 GB of artifact storage a month. An evidence event is one queryable result we keep — roughly one per assertion your tests check; a typical CI run is a few hundred, so a pack is thousands of runs of history. Packs pool across your whole workspace, and one pack covers both events and artifacts — you don't track two meters. Heavy CI can add packs at ~$10 each.

1 pack / seat (1M events + 5 GB)

Each editor bundles a month of evidence events and artifact storage — thousands of runs of cross-run history, regression, and flaky tracking.

Add packs at ~$10 each

Past your capacity, runs keep their status and summary but stop saving detailed evidence until next month — add a pack to keep it all. Opt-in, never an auto-charge.

Artifacts roll off by retention

Request/response bodies and traces count against your pack's storage, then prune at your retention window — no separate per-GB bill.

Example — a 5-developer team: 5 × $12 = $60/mo, including 5 packs (~5M events + ~25 GB) and 1,500 agent credits. Heavy CI past that adds packs at ~$10 each. Illustrative — Cloud pricing is in beta.

Agent credits

The agent is metered in credits.

Each plan includes a pool of agent credits — and on Pro the pool grows with every editor seat. 1 credit = 1¢, billed by what the model actually uses. When the pool runs out the agent pauses until you top up or bring your own key — never an automatic charge.

Cost per question (typical)

GPT-5.4-mini

balanced default

~2 credits

Sonnet 4.6

deeper analysis

~7 credits

Opus 4.8

hardest reasoning

~12 credits

Need more? Top up anytime

$10

500 credits

$25

1,500 credits

+20%

$50

3,500 credits

+40%

Or bring your own key

Add your own Anthropic or OpenAI API key on any paid plan and the agent runs unlimited — zero credits consumed, your choice of model.

Start free. Add a seat when your team comes along.

30 days of Pro on every new workspace · no card required