Back to vendors
K

Kestra

Also known as: Kestra Technologies, Kestra.io

Visit site
Agent infrastructureindependentVerified 2026-07-07

Open source, event driven orchestration platform that unifies data, AI, infrastructure, and business workflows behind a declarative YAML interface, with AI native agents, human in the loop approvals, and more than one thousand two hundred plugins.

Kestra is an open source orchestration platform founded in 2021 by CEO Emmanuel Darras and CTO Ludovic Dehon, built around a declarative, language agnostic approach to workflow automation. Where incumbent orchestrators like Apache Airflow are Python centric and demand specialized data engineering skills, Kestra defines workflows in simple YAML that can be written from a built in code editor or a no code UI, versioned in Git, and deployed through CI/CD. The company raised a twenty five million dollar Series A in March 2026 led by RTP Global, with Alven, ISAI, and Axeleo participating, bringing total funding to thirty six million dollars, and reports more than two billion workflows executed in 2025 across thirty thousand plus organizations including Apple, Toyota, JPMorgan Chase, Bloomberg, Deutsche Telekom, and BHP.

The platform unifies scheduled and event driven automation in one engine, triggered by cron, events, webhooks, or messages, and extends across more than one thousand two hundred plugins covering databases, cloud services, dbt, Spark, Terraform, Ansible, and messaging systems. Kestra 1.0, the long term support release, positions the product as declarative agentic orchestration: AI native agents can be powered by OpenAI, Google Gemini, Anthropic Claude, Mistral, Bedrock, Vertex AI, DeepSeek, or Ollama, combining memory with tools such as web search, code execution, and MCP clients, while a Copilot generates workflow YAML from natural language. Every agent action stays version controlled and auditable, with human in the loop approvals available at critical steps. Enterprise Edition adds SSO, RBAC, multi tenancy, audit logs, encrypted secrets, worker groups, and high availability, deployable on any cloud, on premises, or in air gapped environments, and a fully managed Kestra Cloud with usage based pricing is rolling out alongside the announced Kestra 2.0 distributed execution engine.

Kestra fits engineering and platform teams that want one governed control plane for data pipelines, infrastructure automation, and agentic workflows, especially in hybrid or regulated environments that need self hosting. It is a weaker fit for buyers wanting a turnkey business agent out of the box, since Kestra is the orchestration layer agents run on rather than a prebuilt vertical agent, and paid tier rates are not published.

Vendor details

Canonical URL

https://kestra.io

Category

Agent infrastructure

Subcategory

Declarative agentic orchestration platform

Funding status

Independent, founded in 2021 in Paris by CEO Emmanuel Darras and CTO Ludovic Dehon. Raised a twenty five million dollar Series A on March 31, 2026 led by RTP Global with participation from Alven, ISAI, and Axeleo, bringing total funding to thirty six million dollars after a three million dollar pre seed in 2023 and an eight million dollar seed in September 2024. Reports enterprise revenue growth of twenty five times since the seed round, more than two billion workflows executed in 2025, and customers including Apple, Toyota, JPMorgan Chase, Bloomberg, SoftBank, Deutsche Telekom, and BHP.

Company status

independent

Use cases & customers

Primary use cases

Orchestrating agentic AI workflows with governance and auditabilityData pipeline orchestration across ETL, ELT, and dbtInfrastructure automation spanning Terraform, Ansible, and CI/CDEvent driven and scheduled business process automation

Target customers

Platform and data engineering teamsEnterprises unifying data, AI, and infrastructure automationRegulated and hybrid environment operatorsTeams replacing Airflow or fragmented schedulers

Deployment options

CloudSelf-hostedOn-premisesAir-gapped

Integrations

More than one thousand two hundred plugins spanning databases, cloud storage, APIs, dbt, Airbyte, Spark, Terraform, Ansible, Docker, Kubernetes, and messaging systems, plus an API first design with Python, JavaScript and TypeScript, Java, and Go SDKs. Agents connect to tools including web search, code execution, and MCP clients, and workflows integrate with Git and Terraform for version control and CI/CD.

In practice

A platform team runs data pipelines, infrastructure jobs, and AI agents through separate schedulers and scripts. Kestra unifies them in one declarative control plane where every workflow is versioned, reviewed, and auditable.

An enterprise wants AI agents in production but compliance demands oversight. Kestra runs agents with any major LLM, keeps every action version controlled, and pauses workflows for human approval at critical steps.

A regulated company cannot send workloads to a vendor cloud. Kestra deploys self hosted on Kubernetes or fully air gapped, with RBAC, audit logs, and worker groups isolating sensitive workloads.

Capability coverage

12.0 / 14 capabilities · 86%

Integrations & Tool CallingMore than one thousand two hundred plugins across databases, cloud, dbt, Spark, Terraform, and messaging, and agents combine tools including web search, code execution, and MCP clients, kestra.io homepage and 1.0 release page retrieved 2026-07-08 Full
Workflow OrchestrationCore product is an event driven orchestration engine unifying scheduled and event driven workflows with retries, backfills, subflows, conditional branching, and error handling, kestra.io homepage and GitHub readme retrieved 2026-07-08 Full
Knowledge Grounding & RAGRAG pipelines, evaluation, and retraining are orchestrated with governance and observability, and agents use retrieval tools, but a first class built in grounding or knowledge base layer is not documented, kestra.io homepage and 1.0 page retrieved 2026-07-08 Partial
Human Oversight & GuardrailsHuman in the loop approvals at critical workflow steps are documented, with pause and resume functionality letting humans validate AI generated outputs, kestra.io 1.0 page, docs, and why kestra page retrieved 2026-07-08 Full
Security, Identity & GovernanceEnterprise Edition documents SSO, RBAC, multi tenancy, tenant isolation, encrypted secrets, secrets manager integrations, and audit logs, kestra.io enterprise docs and pricing page retrieved 2026-07-08 Full
Observability & AuditabilityReal time executions, logs, metrics, dependency maps, and flow level SLAs are documented, with audit logs and full auditability of agent actions, kestra.io 1.0 page and enterprise docs retrieved 2026-07-08 Full
Memory & State PersistenceThe engine is designed for long running stateful workflows with retries, backfills, and failure handling built in, and AI native agents combine memory with tools, kestra.io pricing page and 1.0 page retrieved 2026-07-08 Full
Deployment & Data ResidencySelf hosting on Docker or Kubernetes, deployment to any cloud, on premises, hybrid, and air gapped environments are documented first class, kestra.io homepage, pricing page, and 1.0 page retrieved 2026-07-08 Full
Prebuilt Agents, Templates & PacksA blueprints library of real workflow examples is documented, with custom blueprints available in Enterprise Edition and hundreds of built in plugins enabling no code task automation, kestra.io blog on orchestration platforms and why kestra docs retrieved 2026-07-08 Full
Triggers & Channel CoverageCron schedules, events, webhooks, messages, and API triggered executions are documented as one engine to run them all, with event driven and scheduled workflows unified, kestra.io homepage and GitHub readme retrieved 2026-07-08 Full
Model Flexibility & RoutingAI native agents are documented as powered by OpenAI GPT, Google Gemini, Anthropic Claude, Mistral, Bedrock, Vertex AI, DeepSeek, and Ollama, giving customers model choice, kestra.io 1.0 page retrieved 2026-07-08 Full
APIs, SDKs & MCP ExtensibilityAPI first design with Python, JavaScript and TypeScript, Java, and Go SDKs, MCP client support in agents, custom plugin development, and Git and Terraform integrations, kestra.io 1.0 page and GitHub readme retrieved 2026-07-08 Full
Testing, Debugging & OptimizationSyntax validation, autocompletion, faster failure debugging through built in observability, and orchestrated evaluation steps for AI pipelines are documented, but a first class agent testing and optimization suite is not, kestra.io pricing page and homepage retrieved 2026-07-08 Partial
Browser & Computer UseNo browser or computer use capability is described on retrieved pages, kestra.io homepage, 1.0 page, and docs retrieved 2026-07-08 Unable to verify

Recent platform changes

No recent material changes tracked yet.

Pricing

Free open source edition; Enterprise Edition license and Kestra Cloud are quoted through sales with no published rates

license fees on top of a free open source core, scoped by organization size and features, plus usage based pricing on the managed Kestra Cloud

Contact onlyMedium variable costFree tier

What is public

The tier structure is public: a free open source edition, a self hosted Enterprise Edition adding SSO, RBAC, multi tenancy, audit logs, worker groups, and high availability, and a fully managed Kestra Cloud. The company FAQ states Enterprise fees depend on organization size and features, but no dollar rates are retrievable from public pages.

Billing mechanics

Switching from open source to Enterprise Edition involves adding a license key to the configuration and restarting the server, per the company pricing page. Enterprise Edition includes standard support by default with paid upgrades for faster SLAs. Kestra Cloud is operated and maintained by Kestra with usage based pricing announced but not published.

Cost watchouts

Self hosted deployments carry your own infrastructure cost for servers, Postgres or Kafka and Elasticsearch backends, and scaling, which sits outside the license fee.

Variable cost rationale

The open source core is free to run on your own infrastructure, so cost there is your own compute and hosting. Enterprise Edition license fees depend on organization size and the features required, per the company FAQ, and the managed Kestra Cloud is rolling out with usage based pricing announced in the Series A, so cost follows workload once on the managed service.

Additional watchouts

Kestra Cloud pricing is described as usage based but no rate card was retrievable, and the Series A announced its broader rollout, so confirm current Cloud rates directly. Worker Groups are documented as Enterprise Edition only and not yet available in Kestra Cloud, which matters if workload isolation drove the purchase.

Sales call required

Mixed (some tiers require a call)

Free / trial

The open source edition is free forever, self hosted on Docker or Kubernetes

Key ambiguities

How Enterprise Edition license fees scale with organization size, and what the usage based Kestra Cloud rates will be at general availability.

Missing data

No public dollar rates for Enterprise Edition licenses or Kestra Cloud, and no published minimums or contract terms were retrievable.

Verified 2026-07-08

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.