AI Copilot
Describe what you want to orchestrate in natural language, and Copilot generates declarative YAML code you can refine, version, and deploy.
Build and scale workflows faster with AI-assisted code, autonomous agents, and production-ready features. Scale AI workflows in days instead of weeks with full governance and no vendor lock-in.
LTS Release
Apache 2.0 License
900+ Plugins
Deploy anywhere
Trusted by global Enterprises to run mission-critical workflows
Orchestration has been a patchwork: fragile schedulers, vendor lock-in, hand-rolled scripts, and black-box platforms that slow delivery and break governance.
Kestra 1.0 (LTS) unifies data, AI, infrastructure, and business operations under one declarative control plane. Tell the system what you want; Kestra figures out the how, with full auditability and human-in-the-loop approvals.
Describe what you want to orchestrate in natural language, and Copilot generates declarative YAML code you can refine, version, and deploy.
AI-native agents powered by OpenAI GPT, Google Gemini, Anthropic Claude, Mistral, Bedrock, Vertex AI, DeepSeek, and Ollama, combining memory with tools like Web Search, Code Execution, MCP clients, and File System ops to adapt, loop, and orchestrate until goals are met, all fully observable and governed.
Connect Kestra to AI IDEs & agents frameworks via the Model Context Protocol; manage flows and executions from your AI workspace.
Get answers to your questions instantly.
Predictable upgrades, backward compatibility, RBAC, encrypted secrets, tenant isolation proven at scale.
Our extensive ecosystem spans data, AI, and infrastructure. Plugin Versioning allows you to pin and run multiple versions simultaneously for safer, more controlled rollouts.
Reduce friction in complex builds by testing steps independently while keeping production runs clean and consistent.
Enforce reliability standards with built-in SLAs that automatically handle long-running or failed executions to keep workflows compliant and predictable.
Full visibility into complex workflows with a scalable, intuitive dependency view that makes it easy to trace, explore, and navigate data flows at any scale.
Why Use Declarative Orchestration For AI Agents
Describe the desired outcome and constraints; let agents handle the execution details.
Agents can choose the next best action dynamically, while the declarative model ensures repeatable version-controlled code.
Every decision is logged and observable, so you keep oversight without limiting your iteration speed.
CEO & Co-Founder
Replace Fragmented Tools
With A Single Unified Platform
Job schedulers, complex workflow frameworks, custom scripts, and manual glue code spread across multiple systems.
One platform to define, run, and observe workflows, react to event triggers, manage human-in-the-loop approvals, secrets, audit trails, and SLAs.
Fewer moving parts
Reduce tool sprawl and operational overhead.
Lower infrastructure costs
Consolidate systems and cut redundant infrastructure.
Faster delivery
Move from prototypes to production without friction.
Consistent governance
Enforce versioning, auditing, and compliance by default.
100% open source under Apache 2.0 License.
Deploy to any cloud, on-prem or air-gapped environments.
YAML, SDKs (Python, js/TS, Java, Go), or no code UI - all unrestricted and CICD-ready.
RBAC, tenant isolation, encrypted secrets, audit logs.
Horizontal workers, hot reload, safe rollouts, built-in plugin ecosystem.
Real-time executions, logs, metrics, dependency maps, flow-level SLAs.
What could Declarative Agentic Orchestration can do for you?
Book Your Strategy Session