durable-execution
(4)
Temporal Technologies
Temporal Technologies builds a durable execution platform that lets developers write reliable, long-running applications as if failures never happen. Born from the open-source Temporal engine created by ex-Uber and AWS engineers, it automatically persists application state, retries failed steps, and resumes workflows after crashes. Offered as open source and as the managed Temporal Cloud, it is used by more than 3,000 paying customers including Netflix, Snap, Stripe, and Nvidia, and is increasingly the backbone for reliable, stateful AI agents.
Inngest
Inngest is a durable execution platform that gives developers reliable background jobs, workflows, and AI-agent orchestration without managing queues or infrastructure. Founded in 2021 by Tony Holdstock-Brown and Dan Farrelly, the San Francisco company combines event streams, queues, and durable step functions into SDKs for TypeScript, Python, and Go, plus AgentKit for multi-agent systems. It has raised roughly $30 million, most recently a $21M Series A led by Altimeter in 2025, and processes over 100 million executions a day for teams including SoundCloud, Resend, and Replit.
Orkes
Orkes is an enterprise workflow orchestration company built by the creators of Netflix's open-source Conductor engine. It offers a fully managed, cloud-hosted platform that lets developers build, run, and scale durable workflows across microservices, APIs, and - increasingly - AI agents, giving companies a reliable way to put long-running, event-driven, and agentic processes into production without reinventing the underlying orchestration plumbing.
DBOS, Inc.
DBOS, Inc. is a San Francisco- and Cambridge-based developer infrastructure company commercializing durable execution technology born from a joint MIT-Stanford research project led by Postgres creator Michael Stonebraker and Apache Spark creator Matei Zaharia. Its flagship open-source library, DBOS Transact, adds crash-proof workflow orchestration and built-in observability to ordinary Python and TypeScript programs by persisting application state in Postgres, so software resumes exactly where it left off after a failure - without a separate workflow server. The company pairs the library with DBOS Conductor and DBOS Cloud for managing and hosting reliable backends, AI agents, and long-running workflows.