Temporal Workflows
by Temporal
Build applications that never lose state, even when everything else fails. Temporal is an open-source durable execution platform that lets developers write long-running, fault-tolerant workflows as ordinary code. Founded in 2019 by Maxim Fateev and Samar Abbas, the same team that built Uber's Cadence orchestrator, Temporal turns transient business processes (order fulfilment, payment retries, multi-step ML pipelines) into deterministic state machines whose state survives process crashes, deploys, and full datacentre failures.
Performance Scores
3 rankings evaluated
Score range: 8.4 – 9.1
-
#1Best Durable Workflow Engines for Production in 2026
Score: 8.7 · Best for: Engineering teams building long-running, fault-tolerant workflows in code at SaaS and fintech scale
-
#1Best Open-Source Workflow Engines for Engineers in 2026
Score: 8.4 · Best for: Engineering teams building mission-critical distributed workflows that must survive multi-day delays and infrastructure failure
-
#1Best Process Orchestration Platforms 2026
Score: 9.1 · Best for: Engineering teams building workflows that must survive infrastructure failure
Key Facts
pricing
| Attribute | Value | As of | Source |
|---|---|---|---|
| Actions pricing (May 2026) | Consumption is metered in Actions at about $50 per million ($0.00005 each); volume discounts begin at 5M Actions across all namespaces | May 2026 | Temporal Cloud pricing documentation |
| Temporal pricing (July 2026) | Temporal pricing re-verified September 2026: self-hosted remains free (MIT). Temporal Cloud's entry tier changed on 2026-09-14 from 'Essentials' ($100/mo minimum, 5% of usage) to a no-base-fee 'Pay-As-You-Go'/'Developer' plan billed at 10% of usage spend; Actions remain about $50/million, volume-tiered. Business remains from $500/mo (greater of $500/mo or 10% of usage; 2.5M Actions), and Enterprise/Mission Critical remain custom, billed annually. | Sep 2026 | Temporal blog: "Pay only for what you use" (Sept 14, 2026) |
| Self-hosted licensing (May 2026) | Self-hosted Temporal is free and open-source under the MIT license; the user provides and operates the cluster and its database | Sep 2026 | temporalio/temporal GitHub repository LICENSE file |
| Temporal Cloud plans (May 2026) | Temporal Cloud plans (as of September 2026): Pay-As-You-Go / Developer support has no base monthly fee, billed at 10% of usage spend, with Actions at $50 per million; Business starts at $500 per month (2.5M Actions), billed as the greater of $500/mo or 10% of usage spend; Enterprise and Mission Critical are custom-priced (10M Actions included) and billed annually. The $100/month Essentials plan is no longer listed on Temporal's pricing pages, following a September 14, 2026 change introducing no-minimum consumption pricing for the entry tier. | Sep 2026 | Temporal Cloud Pricing (temporal.io/pricing.md) |
| Cloud minimum | $100 per month minimum (Essentials plan; billed the greater of plan fee or 5%-10% of usage), consumption billing on actions and storage | Jul 2026 | Temporal pricing page |
General
| Attribute | Value | As of | Source |
|---|---|---|---|
| Durable Execution | Workflows automatically capture state at every step and recover from failures without data loss. | Jul 2026 | Temporal — How It Works |
| Built-in Retries & Timeouts | Activities automatically retry on failure with configurable timeouts and recovery logic. | Jul 2026 | Temporal Docs |
| Event History & Retention | Active storage for open workflows and retained storage for historical events. | Jul 2026 | Temporal Pricing |
| Startup Program | Startups under $30M in funding receive $6,000 in free Temporal Cloud credits. | Jul 2026 | Temporal Pricing |
Core Features
| Attribute | Value | As of | Source |
|---|---|---|---|
| Action-Based Pricing | Billing is based on actions, with tiered pricing starting at $50 for next 5M actions. | Jul 2026 | Temporal Pricing |
Limits & Quotas
| Attribute | Value | As of | Source |
|---|---|---|---|
| Capacity Scaling | On-demand mode starts at 500 APS, with optional provisioned capacity using TRUs. | Jul 2026 | Temporal Pricing |
Support
| Attribute | Value | As of | Source |
|---|---|---|---|
| Support Tiers | Support ranges from 1 business day to 24/7 enterprise support with P0 response times. | Sep 2026 | Temporal Cloud Pricing (temporal.io/pricing.md) |
| Multi-Cloud & Multi-Region | Available with cloud platform support across regions and multi-cloud environments. | Jul 2026 | Temporal Pricing |
community
| Attribute | Value | As of | Source |
|---|---|---|---|
| GitHub stars | 21,900+ stars | Jul 2026 | GitHub API |
technical
| Attribute | Value | As of | Source |
|---|---|---|---|
| Official SDKs | 8 official SDKs (Go, Java, Python, TypeScript, .NET, PHP, Ruby, Rust) | Jul 2026 | Temporal docs (Develop section) |
Strengths
- ●History-replay model gives effectively exactly-once execution without explicit checkpoints
- ●First-party SDKs in Go, Java, TypeScript, Python, .NET, PHP, and Ruby
- ●Public production deployments at Stripe, Snap, Coinbase, and Netflix
- ●Permissive MIT licence on the OSS edition; Cloud and self-host both viable in production
- ●Deterministic replay guarantees exactly-once semantics across crashes and restarts
- ●Six officially supported SDKs (Go, Java, Python, TypeScript, .NET, PHP) with full feature parity
- ●Production-proven at Netflix and Stripe scale — over 1 billion workflow executions per month at large deployments
- ●Active development with weekly commits and a dedicated company (Temporal Technologies) behind the project
- ●Durable execution with saga compensation as a first-class SDK construct
- ●Eight SDKs with replay-based determinism testing and a time-skipping test environment
- ●Workflow definitions live in version-controlled source rather than in a GUI
- ●Multi-region replication and Worker Versioning reached GA in May 2026
Limitations
- ●Steep conceptual learning curve — workflows, activities, signals, queries, replay
- ●Postgres or Cassandra cluster operationally non-trivial at high throughput when self-hosted
- ●Visual UI is a debugging surface, not an authoring tool — code-first only
- ●Operational complexity is high — requires Cassandra or PostgreSQL, history service, matching service, and worker processes
- ●Workflows must be written as deterministic code, which constrains how external APIs and randomness are called
- ●Learning curve is steep for teams unfamiliar with event-sourced systems
- ●Hard ceilings force Continue-As-New: 51,200 events and 50 MB per workflow, 1,000 children per parent
- ●No built-in alerting; Prometheus or Datadog must be wired up separately
- ●Cloud history retention caps at 90 days and self-hosted Archival is still experimental
- ●Self-hosting means four services plus a database plus Elasticsearch
- ●No connector catalogue; every integration is code the team writes and maintains
Based on evaluations in 3 rankings: Best Durable Workflow Engines for Production in 2026, Best Open-Source Workflow Engines for Engineers in 2026, Best Process Orchestration Platforms 2026
Pricing
Temporal Workflows offers multiple products and plan tiers whose usage allowances vary by plan; the vendor's pricing page is the authoritative reference for current plans and limits.
View official pricing →About Temporal Workflows
Temporal is an open-source durable execution platform that lets developers write long-running, fault-tolerant workflows as ordinary code. Founded in 2019 by Maxim Fateev and Samar Abbas, the same team that built Uber's Cadence orchestrator, Temporal turns transient business processes (order fulfilment, payment retries, multi-step ML pipelines) into deterministic state machines whose state survives process crashes, deploys, and full datacentre failures.
The architecture separates workflow logic from infrastructure. Application code defines workflows and activities in any of seven supported SDKs (Go, Java, Python, TypeScript, .NET, PHP, Ruby as of Q2 2026), while a Temporal cluster persists every step to Cassandra, MySQL, PostgreSQL, or Elasticsearch. If a worker dies mid-execution, another worker resumes the workflow from the last completed step using event-history replay; developers write code as if failures cannot happen, and the runtime makes that abstraction true.
Temporal Cloud, the company's managed offering, launched general availability in late 2022 and as of July 2026 is priced at a $100/month minimum (Essentials plan) with consumption-based billing on actions, active workflows, and storage retention. Self-hosted deployments under the MIT licence remain free, and the open-source repository on GitHub passed 13,800 stars in May 2026. Production users include Snap, Box, Stripe, HashiCorp, Coinbase, Datadog, and Netflix; the company reports 2,000+ companies running Temporal workflows in production.
Where Temporal differs from BPMN engines (Camunda) and DAG schedulers (Airflow, Prefect) is the durable-execution primitive: the entire workflow lives in code, history is the source of truth, and signals or queries can mutate or inspect a running workflow without restarting it. The trade-off is operational weight; a Temporal cluster is several stateful services and a database, which is overkill for cron-style jobs but proportionate for systems where a lost workflow costs real money.
flowchart LR
A[Client SDK] -->|Start workflow| B[Frontend Service]
B --> C[(History DB)]
B --> D[Matching Service]
D --> E[Worker Pool]
E -->|Activities| F[External APIs]
E -->|Replay events| C
C --> G[History Service]
G --> B
Editor's Note: Across 7 ShadowGen Temporal deployments in 2025–26, median time-to-first-production-workflow was 11 days for teams already comfortable with Go or Python, and 26 days for teams coming from Airflow. The bill that surprises people is storage retention: a 90-person fintech client carried 480 GB of event history after 14 months on default retention; setting per-namespace retention to 30 days dropped that to 41 GB and saved roughly $190/month on Cloud. Where it breaks: workflow versioning is unforgiving if you skip GetVersion() calls during a deploy that touches in-flight workflows. Plan a determinism-test step in CI before the second production release. — Rafal Fila, ShadowGen
Integrations (4)
Other Workflow Automation Tools
Activepieces
No-code workflow automation with self-hosting and AI-powered features
Workflow AutomationAutomatisch
Open-source Zapier alternative
Workflow AutomationBardeen
AI-powered browser automation via Chrome extension
Workflow AutomationCalendly
Scheduling automation platform for booking meetings without email back-and-forth, with CRM integrations and routing forms for lead qualification.
Workflow AutomationSee How It Ranks
Best Automation Platforms for AI Orchestration 2026
This ranking answers one question: how many real business applications can an AI agent act on out of the box? It evaluates nine platforms as of August 2026 on the reach they give an agent, not on the workflow logic they can express. That boundary is deliberate, because two neighbouring pages on this site answer different questions. Best Process Orchestration Platforms 2026 scores multi-step process control, error handling and state management. Best AI Agent Platforms 2026 scores building and hosting the agent itself. This page scores the layer between them: the connective tissue that lets an agent already built elsewhere reach the applications a business actually runs on. A platform that leads one of those pages can place low here, and two of them do. Scores derive from application and action catalogue counts, the exposure model each platform uses to publish those catalogues to an agent, setup effort, failure handling and cost per agent action. Every figure was retrieved from a vendor-owned surface on 11 August 2026 unless an earlier date is stated against it.
Best Durable Workflow Engines for Production in 2026
A ranked list of the best durable workflow engines for production deployments in 2026. Durable workflow engines persist execution state to a database so that long-running workflows survive process restarts, deployments, and infrastructure failures. The ranking covers Temporal, Prefect, Apache Airflow, Camunda, Windmill, and n8n. Tools were evaluated on production reliability, developer experience, scalability, open-source health, and documentation quality. The shortlist intentionally mixes code-first engines (Temporal, Prefect, Airflow) with hybrid visual platforms (Camunda, Windmill, n8n) to reflect how production teams actually choose workflow engines in 2026.
Questions About Temporal Workflows
How much does Kestra cost in 2026?
Kestra's core platform is free and open-source (Apache 2.0) with unlimited flows and executions when self-hosted. As of July 2026 the Enterprise Edition is a quote-based annual subscription priced per instance, and the managed Kestra Cloud is in early access with usage-based pricing; neither publishes dollar prices, so budgeting starts with infrastructure costs plus a sales conversation.
What is QuantumBPM?
QuantumBPM is a combined BPMN 2.0 process engine and DMN 1.5 decision engine launched in 2026 by Coroid s.r.o., a Slovak company founded by Richard Bízik. Its decision engine passed 3,367 of 3,391 tests (99.29%) in the official DMN Technology Compatibility Kit in February 2026, listed as QuantumDMN Engine 1.0.0. The platform runs as managed SaaS or self-hosted on Docker/Kubernetes, ships MIT-licensed SDKs for Go, Java, Python, and JavaScript, and uses credit-based pricing: Starter $199/month and Scale $1,199/month as of July 2026, with custom self-hosted/enterprise licensing.
Can you self-host Trigger.dev in 2026?
Yes. Trigger.dev is open-source under the Apache 2.0 license and supports self-hosting via Docker Compose or Kubernetes. Self-hosting gives you full control over your job runs and data, but you take on operating Postgres, Redis, the worker queue, and the dashboard yourself.
What is the difference between Inngest and traditional cron for job scheduling?
Inngest is a durable event-driven job platform with retries, observability, and step functions, while traditional cron is a time-based scheduler with no built-in retry, queueing, or observability. Inngest is closer to a job queue than to cron, and is most useful for application background work that needs reliability guarantees.
Learn More
Which Automation Platforms Can an AI Coding Agent Actually Edit?
This guide compares the canonical workflow representation of ten business process automation platforms and asks a single practical question of each: can an AI coding agent read that representation, change it, and put it back? It covers what each platform exports, what the export silently drops, which route requires a paid or enterprise tier, and where the round trip breaks. Verified against vendor documentation on 26 August 2026.
AI Coding Agents and Automation Platforms: Three Integration Models
This guide classifies the ways an AI coding agent can connect to a business process automation platform into three integration models: reference servers that teach an agent a platform's schema, control-plane servers that let it read and edit workflows on a live instance, and action-proxy servers that expose the platform's app catalogue as callable tools. It compares what eight platforms shipped as of 26 August 2026, with the documented endpoint, the authentication method, and the billing unit each model consumes.
QuantumBPM vs Camunda 2026: Single-Binary Challenger vs the BPMN Incumbent
QuantumBPM (launched 2026, Coroid s.r.o., Slovakia) packages a BPMN 2.0 runtime and DMN 1.5 decision engine into one Go binary backed by Temporal and PostgreSQL. Camunda (Berlin, founded 2013) is the category incumbent: Camunda 7 (Apache 2.0, in maintenance) and the Zeebe-based Camunda 8 platform. This comparison covers product structure, architecture, DMN TCK conformance with recording dates, deployment, pricing, and vendor maturity, verified July 2026.