Free plan with $5/month credits, Hobby ($10/month with $10 credits), Pro ($50/month with $50 credits), Enterprise with custom pricing. Billed by compute usage and run invocations.
Not given
Key features
AI agents that survive refreshes, redeploys, and crashes
Long-running tasks without timeouts
Automatic retries with configurable backoff
Fair queuing with concurrency controls
Real-time observability and monitoring
Human-in-the-loop workflows
Streaming LLM responses to frontend
Scheduled tasks and cron jobs
Multi-region deployment
Elastic auto-scaling infrastructure
In-app Inbox component with real-time notifications
Multi-channel notifications across email, SMS, push, and chat
User-controlled notification preferences with quiet hours and snooze
Digest and rate limiting capabilities
Single workflow routing across all channels
Full conversation history and context for AI agents
Human-in-the-loop approval before agent actions
Support for multiple AI frameworks including Claude, Vercel AI SDK, and LangChain
Role-Based Access Control (RBAC) and team management
Enterprise compliance with HIPAA, SOC 2, and ISO 27001
What makes it different
Durable execution that survives crashes and redeploys
No infrastructure to manage or timeouts to worry about
Stream responses directly to frontend with built-in Realtime
Full TypeScript support with no runtime restrictions
Unopinionated about AI intelligence - keep your existing models, prompts, and business logic
One API for all channels eliminating need to rebuild agent logic for each channel
Open-source with transparent code and active community
Full message context automatically preserved across conversations
Multiple deployment options - cloud, self-hosted, or open-source