What the core is. The runtime layer everything else runs on — planning, context, governed execution, verification, replay, approvals — plus the Projects control plane you launch/watch/approve missions from. It's open-core: the runtime image is public (ghcr.io/redevops-io/redevops-runtime) and the Helm chart is in the public agentic-os repo. The business apps are layered on top separately (Apps install).
What you need. A Linux, macOS or Windows machine or server where you have administrator rights (root / sudo / local-admin) for the self-host path, or a cloud account with permission to create resources (Kubernetes, a registry, a database) for the cloud path. Docker, package installs and port binding all require admin. This is not a phone/tablet app and not a hosted SaaS you log into.
Questions or stuck? Join the ReDevOps Slack — #core-runtimes is the channel for the core runtime stack (Mission runtime, Context Runtime, permissions, the Projects control plane). We answer there.
The Projects control plane + API ship as one service in the public agentic-os package. Pick your OS.
pip (Python 3.11+ and Git required):
sudo apt install -y python3 python3-venv git
python -m venv .venv && . .venv/bin/activate
pip install 'agentic-os[projects] @ git+https://github.com/redevops-io/agentic-os'
agentic-os-projects # → http://127.0.0.1:8787
Docker (only Docker required):
git clone https://github.com/redevops-io/agentic-os && cd agentic-os
docker build -f Dockerfile.projects -t redevops-projects .
docker run -p 8787:8787 redevops-projects # → http://localhost:8787
pip (Homebrew for prerequisites):
brew install python@3.12 git
python3 -m venv .venv && . .venv/bin/activate
pip install 'agentic-os[projects] @ git+https://github.com/redevops-io/agentic-os'
agentic-os-projects # → http://127.0.0.1:8787
Or Docker — identical to Linux (Docker Desktop for Mac; Apple silicon & Intel both fine).
pip (winget for prerequisites, PowerShell):
winget install -e --id Python.Python.3.12
winget install -e --id Git.Git
python -m venv .venv; .venv\Scripts\Activate.ps1
pip install 'agentic-os[projects] @ git+https://github.com/redevops-io/agentic-os'
agentic-os-projects # → http://127.0.0.1:8787
Or Docker Desktop (WSL2 backend) — run the Docker steps from inside your WSL2 distro so paths stay on the Linux filesystem; identical to the Linux tab.
Run the governed runtime inside Kubernetes. The reference-system/redevops-runtime chart deploys the public runtime; the deployment is itself a mission (a post-install hook runs acceptance checks and fails the release if they don't pass).
git clone https://github.com/redevops-io/agentic-os # self-contained demo (bundled retrieval, no external model): helm install rr ./agentic-os/deploy/reference-system/redevops-runtime --set retriever.backend=inmemory # or point at your model + a managed pgvector DSN: helm install rr ./agentic-os/deploy/reference-system/redevops-runtime \ --set model.endpoint=http://my-llm:8000/v1 \ --set retriever.pgvector.dsn=postgres://user:pass@host:5432/redevops_rag
Verify: kubectl -n <ns> port-forward svc/rr-redevops-runtime 8080:8080 then curl localhost:8080/health → {"status":"ok"}. The full Projects/Sidekick combo (control plane + infra operator + Prometheus/Grafana/Loki) is the detailed quickstart.
Provision managed Kubernetes + a registry + a managed Postgres, then Helm-install the runtime. The chart is provider-agnostic — only model.endpoint and the pgvector DSN change per cloud. Full model wiring per provider: redevops.io/integrations.
aws configure
eksctl create cluster --name redevops --region us-east-1 --nodes 2
aws eks update-kubeconfig --name redevops --region us-east-1
git clone https://github.com/redevops-io/agentic-os
helm install rr ./agentic-os/deploy/reference-system/redevops-runtime \
--set model.endpoint=http://your-openai-compatible-gateway/v1 \
--set retriever.pgvector.dsn=postgres://user:pass@your-rds:5432/redevops_rag
Model: Amazon Bedrock via an OpenAI-compatible gateway. Full AWS integration →
gcloud container clusters create-auto redevops --region europe-west1
gcloud container clusters get-credentials redevops --region europe-west1
git clone https://github.com/redevops-io/agentic-os
helm install rr ./agentic-os/deploy/reference-system/redevops-runtime \
--set model.endpoint=http://your-openai-compatible-gateway/v1 \
--set retriever.pgvector.dsn=postgres://user:pass@your-cloudsql:5432/redevops_rag
Model: Vertex AI via the Google Agent stack. Full Google integration → (GKE nodes: use e2-standard-2+ — system overhead starves an e2-medium.)
az aks create -g redevops-rg -n redevops --node-count 2 --generate-ssh-keys
az aks get-credentials -g redevops-rg -n redevops
git clone https://github.com/redevops-io/agentic-os
helm install rr ./agentic-os/deploy/reference-system/redevops-runtime \
--set model.endpoint=https://your-azure-openai/v1 \
--set retriever.pgvector.dsn=postgres://user:pass@your-azure-pg:5432/redevops_rag
Model: Azure OpenAI. Full Azure integration →
doctl kubernetes cluster create redevops --region nyc3 --count 2
doctl kubernetes cluster kubeconfig save redevops
git clone https://github.com/redevops-io/agentic-os
helm install rr ./agentic-os/deploy/reference-system/redevops-runtime \
--set model.endpoint=http://your-openai-compatible-gateway/v1 \
--set retriever.pgvector.dsn=postgres://user:pass@your-do-db:25060/redevops_rag
Model: any OpenAI-compatible endpoint. Full DigitalOcean integration → (Or a single Droplet + docker compose for a pilot.)
Governed by default. Whether you Helm-install directly or drive it with Sidekick, a cloud deploy is a governed mission — plan → approval gate → provision → verify → saga. Validated on EKS, GKE and DOKS: provision → helm install (public chart + public redevops-runtime:0.3.1 image) → /health 200.