Skip to content

Deterministic execution platform

Build engineering automation once.
Run it consistently everywhere.

Engineering Runtime turns repeatable engineering work into reusable Engineering Capabilities and executes them through one deterministic, policy-governed contract.

Proprietary Runtime. Open capability ecosystem.
The caller changes. The execution contract stays the same.

runtime
# install with Homebrew…
$ brew install engineeringruntime/tap/engineering-runtime

# …or the script — no account required, any machine
$ curl -fsSL https://raw.githubusercontent.com/engineeringruntime/engineering-runtime-releases/main/install.sh | sh

# confirm the Runtime is installed
$ runtime version
runtime <latest version>

# discover the built-in surfaces — no token or account required
$ runtime --help
Available Commands:
  capability  Validate and execute Engineering Capabilities
  files       Execute files operations
  github      Execute github operations
  resolve     Check command resolution with zero side effects

# create a test file in your current directory
$ runtime files write runtime-test.txt 'hello from Runtime'
  • Engineers
  • CI/CD
  • Automation
  • AI agents

A proprietary Runtime. An open capability ecosystem.

Engineering Runtime is not an open-source Runtime with a closed enterprise fork. The Runtime is a proprietary product across its editions. The ecosystem around it is open: teams can create, review, share, and improve Engineering Capabilities without changing the Runtime itself.

Proprietary Runtime

The Runtime source remains private. Public binaries, public documentation, and the no-cost Community edition make the product accessible; they do not make the Runtime open source.

Open capability ecosystem

Capabilities are the extension and participation surface. Publish them for the community, keep them private inside a company, or build reusable packs for a team or customer.

The Runtime stays focused. The capability ecosystem is where reusable engineering knowledge grows.

How capabilities work Read the source-model FAQ


Stop rebuilding the same engineering operation

Engineering automation is usually implemented several times over — once per caller. Each implementation drifts in behaviour, authentication, policy, error handling, and auditability.

Without Engineering Runtime

EngineerCI/CDAutomationAI agent
shell scriptworkflow logicintegrationgenerated commands

Four implementations. Four ways to drift.

With Engineering Runtime

EngineerCI/CDAutomationAI agent
Engineering Capability Engineering Runtime Engineering systems

One implementation. One contract. Every caller.

Define the engineering operation once, document it once, version it, review it, and execute it repeatedly.

Without Engineering Runtime With Engineering Runtime
The same task becomes several scripts and workflows The task becomes one reusable capability
Every caller implements its own execution path Every caller uses the same Runtime contract
AI frequently creates another one-off script AI can discover and invoke an existing capability
Engineering knowledge remains scattered Engineering knowledge becomes versioned and executable
Governance is added differently in every tool Policy is evaluated as part of execution

One execution foundation for engineering work

Reusable capabilities

Capture proven engineering operations as files that can be reviewed, versioned, tested, shared, and improved.

One contract for every caller

Use the same operation from a developer terminal, CI/CD pipeline, automation platform, or AI agent.

Governed execution

Apply context, policy, authentication, execution, and audit through a consistent lifecycle.

Local-first adoption

Run where engineering work already happens. A central service is not required to begin.


Capabilities are executable engineering knowledge

An Engineering Capability represents a reusable engineering operation or workflow — repository health checks, configuration changes, CODEOWNERS updates, file transformations, release preparation, platform operations.

Capabilities turn knowledge that would otherwise live in scripts, wiki pages, CI workflows, AI conversations, and individual experience into a durable unit.

Engineering knowledge Capability Version controlled Documented and reviewed Executable by Human · CI · Automation · AI

A capability that works for one engineer can be executed through the same Runtime by a pipeline or an AI agent.

Explore capabilities


Deterministic by design

Every Runtime operation follows the same lifecycle, regardless of who initiated it.

01BootstrapPrepare the Runtime Home before any work runs.
02ContextResolve where the operation will execute.
03PolicyEvaluate the rules before the operation proceeds.
04AuthenticationUse the platform's own supported authentication.
05ExecutionRun through a provider or engine.
06AuditRecord the outcome.

This creates a clear boundary between the system requesting an operation and the system where that operation is performed.

Explore capabilities Read the vision


Built for AI without depending on one AI vendor

AI assistants are excellent at understanding intent, reasoning, planning, and interpreting results. Engineering Runtime is responsible for deterministic execution.

Your AI assistant

  • Understands intent
  • Reasons and plans
  • Discovers capabilities
  • Supplies inputs
  • Interprets results

Engineering Runtime

  • Applies the execution contract
  • Evaluates policy
  • Uses defined authentication
  • Executes the operation
  • Records the outcome

Any capable AI agent that can understand the documented Runtime contract and invoke the binary can use the same execution model.

Reason with AI. Execute with Engineering Runtime.

Try Engineering Runtime with an AI assistant

Ask an AI assistant to follow the installation documentation, discover an appropriate capability, and run it.

Start with AI


Use curated providers, or the tools your teams already trust

Runtime Providers

Providers expose curated operations for an engineering system through a Runtime-defined interface. Runtime currently ships providers for:

filesgithub

Command Engine

Run approved engineering binaries through Runtime governance instead of recreating every existing tool as a provider.

terraformkubectlgcloudother approved CLIs

Providers are valuable where a stable Runtime abstraction improves the experience. Existing CLIs remain useful where their native interface is the right fit.

Engineering Runtime integrates with engineering systems; it does not attempt to replace them.

View technical documentation


Start locally. Add organisational governance when it is needed.

Runtime execution, policy evaluation, provider execution, the Command Engine, and the installed Runtime contracts all operate locally. Engineering Runtime is useful without a central platform.

For organisations that later require central policy, configuration, Runtime identity, fleet visibility, and management, the optional Control Plane represents the enterprise governance direction. It remains outside the execution path and is not required to start.

See what Runtime offers Explore capabilities


Designed for teams building modern engineering platforms

Platform and DevEx teams

Turn repeatable platform operations into reusable capabilities instead of maintaining separate scripts for every team and workflow.

CI/CD and automation teams

Standardise how engineering operations execute across repositories, pipelines, and environments.

Teams adopting AI agents

Give agents a documented, policy-aware execution boundary rather than relying only on generated commands.

Engineering teams

Discover and reuse proven operations while keeping the underlying engineering platforms authoritative.


Why this matters now

AI has made it easier to generate automation. It has not removed the need for repeatability, ownership, policy, versioning, documentation, authentication, or audit.

For known engineering problems, the best outcome is often not another generated script. It is the discovery and execution of an existing, trusted capability.

Without reusable capabilities

Taskgenerate automationexecutediscardgenerate again

With Engineering Runtime

Taskdiscover capabilityexecute through Runtimereuse and improve

As the capability ecosystem grows, engineers and AI agents can spend less time recreating solved automation and more time applying, improving, and extending shared engineering knowledge.

Start with a real engineering operation

1Install Engineering Runtime 2The Runtime prepares its Home on the first command 3Run a capability from a terminal, pipeline, automation system, or AI assistant

Created by Kishore Babu Gutta

Engineering Runtime is an independent software project created and maintained by Kishore Babu Gutta, a platform engineer focused on deterministic engineering automation, platform engineering, and governed AI-assisted execution.

The project is being built to transform reusable engineering knowledge into Engineering Capabilities that humans, CI/CD systems, automation, and AI agents can execute consistently through one trusted Runtime.

About the creator

Join the Engineering Runtime community

Engineering Runtime is being built around practical engineering automation and reusable capabilities.

Where Purpose
💬 GitHub Discussions Questions, capability ideas, architecture discussions, and feedback
🐞 GitHub Issues Bugs and feature requests
📘 Technical Documentation Installation, providers, capabilities, configuration, and CLI reference
✉️ Contact Product enquiries and support

One execution contract

Intent can start anywhere.
Execution should stay deterministic.

  • HumansReason
  • AI agentsPlan
  • AutomationOrchestrates
  • CapabilitiesEncode the work

One governed path. Engineering Runtime executes.