Flat-rate plans

Forever Agents pricing

Start with a trial, upgrade for the extension and full dashboard access, or go Enterprise for a dedicated instance and custom volume through Google Cloud Marketplace private offers.

Trial

$0/mo

7 days, no credit card

  • 1 seat, 1 concurrent agent
  • 100 task runs total
  • Gemini Flash model
  • Brain memory: 7-day retention
  • Dashboard + full API access
Start trial

Starter

$49/mo

For a solo builder

  • 1 seat, 2 concurrent agents
  • 1,000 task runs / month
  • Gemini Flash model
  • Brain memory: 30-day retention
  • IDE extension + audit API
  • Email support (2 business days)
Start Starter

Builder

$149/mo

For heavy async workloads

  • 1 seat, 5 concurrent agents
  • 5,000 task runs / month
  • Flash + Pro-class models (15% of runs)
  • Brain memory: 90-day retention
  • Priority model fallback routing
  • Email support (2 business days)
Start Builder

Team

$349/mo

Shared workspace

  • 5 seats, 15 concurrent agents
  • 20,000 task runs / month
  • Flash + Pro-class models (8% of runs)
  • Brain memory: 1-year retention
  • Team queues + admin policy controls
  • Priority support (1 business day)
Start Team

Enterprise

$499/mo

Dedicated instance, private offers

  • Unlimited seats and agents
  • 50,000 task runs / month fair use
  • Unlimited Pro-class inference (BYOK)
  • Dedicated Cloud Run instance + isolation review
  • Named contact, 4h sev-1 response
  • Custom volume via Marketplace private offer
Contact sales

What counts as a task run?

A task run is one claimed task executed end-to-end by an agent: claim, model calls, tool calls, receipt write, completion. Retries after infrastructure failure do not count twice. If you hit your monthly limit, queued tasks pause until the next cycle or an upgrade — nothing is dropped. Model usage is included in the flat rate: Trial and Starter run on Gemini Flash; Builder and Team include a Pro-class allowance (15% and 8% of monthly runs respectively — beyond that, runs continue on Flash, never blocked). Enterprise includes unlimited Pro-class inference when you connect your own Google Cloud project (BYOK) — inference bills to your GCP account at Google's rates, using your existing commit. Without BYOK, Enterprise includes a 5% Pro-class allowance on platform billing. Custom volume and terms are available through Google Cloud Marketplace private offers.

Enterprise deployment

Enterprise runs as a dedicated, single-tenant instance of the Forever Agents platform on Google Cloud — isolated Cloud Run services and a dedicated Firestore namespace, with your data-retention policy and an isolation review as part of onboarding. Connect your own Google Cloud project via BYOK (service-account impersonation, no keys shared, revocable on your side at any time) and all Vertex AI inference bills directly to your GCP account — counting toward your Google Cloud commit. Purchase directly or through a Google Cloud Marketplace private offer for custom volume and terms.

Available on Google Cloud MarketplaceTermsPrivacy