The Enterprise Agent Harness.
Nightshift gives agent operators the choice for where compute, storage, data, and models live. Fully open-source, Apache 2.0.
Chat your team already knows.
Feature parity with ChatGPT or Claude. Open Nightshift, start a conversation, and agents work end to end on your infrastructure.
- Plain-English prompts that kick off multi-step agent runs
- Real artifacts inline: dashboards, reports, code, alerts
- One-shot or on any recurring schedule
Agents and skills, made easy
Build an agent for every job: Purchasing, Accounting, a Compliance monitor. Each carries its own prompt and scope. Layer in skills as reusable playbooks your team writes once and agents apply on every run.
- Skills are markdown specs your team reads, reviews, and versions
- Apply skills globally or scope them to a single agent
Real work, not just chat
Dashboards, reports, spreadsheets, code, apps. Every artifact an agent produces is browsable, shareable, and traceable back to the run that made it.
- Native types: Documents, Spreadsheets, Apps, Code, Images
- Pin to dashboards, share with your team, or keep private
- Every artifact links back to the run, the agent, and the prompt
Plug into the tools you already use
A growing library of first-party connectors into the tools your team lives in. We ship new ones every release, and the integration framework is open so you or the community can add your own.
- OAuth or service-account auth: your IdP, your scopes
- Read and write: agents reply in chat, open PRs, update tickets, query warehouses
- Scoped per agent, audited per run, revocable any time
More than an app.
Nightshift is a Kubernetes-native platform underneath. Kernel isolation, eBPF networking, end-to-end observability, and one Helm chart for the whole stack.
Kernel-level enforcement
Nightshift leverages Cilium and Tetragon for kernel-level enforcement. Agent pods run inside Kata micro-VMs with their own guest kernel, Cilium filters every packet leaving the pod via eBPF, and Tetragon hooks syscalls for runtime policy enforcement.
Cilium
First line of defense. eBPF network policies block agents from reaching anything you haven't explicitly allowed, enforced inline in the kernel before a single packet leaves the pod.
Tetragon
If traffic gets past the network layer, Tetragon hooks syscalls at the kernel level via kprobes. Malicious behavior is blocked before it ever touches userspace.
Kata Containers
Last resort. Even if every other defense is bypassed, the agent is trapped inside a hardware-isolated micro-VM with its own guest kernel. A kernel exploit can't escape to the host.
Tetragon
If traffic gets past the network layer, Tetragon hooks syscalls at the kernel level via kprobes. Malicious behavior is blocked before it ever touches userspace.
nightshift-api · REST
Operators query any running agent on demand.
Running Agents
each pod exposes metrics + logs at the kernel
Tetragon / Cilium
kernel-level event and metric pipeline
Prometheus
metrics time-series
Loki
log aggregation
Grafana
operator dashboards · alerts · OpenTelemetry
Full visibility into every agent
Query logs, metrics, and processes from any running agent via REST API. Nightshift captures kernel-level events through Tetragon and Cilium and ships everything to Grafana, Prometheus, and Loki for your operator team.
One Helm chart. Any cluster.
The full stack on any cluster in seconds: API, Agent Worker, Storage, Auth, and the Nightshift UI. Same chart deploys to your production cluster, cloud, on-premise, or fully air-gapped.
# Spin up the full stack on a local cluster
make eks-quickstart
# Port-forward and open your browser
kubectl -n nightshift port-forward
svc/nightshift-nightshift-ui 13000:3000
→ http://localhost:13000Why Nightshift?
Enterprise-Ready
Built for the controls enterprise security teams expect: hardened defaults, IDP-backed SSO, and an architecture that scales with your org.
Agent Agnostic
Run any agent SDK (Claude, OpenAI, LangGraph, or your own) on the same runtime without lock-in.
Easy to Deploy and Manage
One Helm chart for the full stack. Same chart deploys to a local cluster, your production cluster, on-premise, or fully air-gapped.
Multi-Tenant Support
Strong tenant isolation for teams, business units, or customers — namespaces, identities, and policies scoped per tenant.
Feature-Rich
UI feature parity with the chat products your team already uses, so adoption is one conversation, not a rewrite.
Sovereign
Your compute, your data, your models. Deploy Nightshift on infrastructure you control — cloud, on-premise, or fully air-gapped.
Latest from the community
Shipping Real Apps with Harness Engineering
Dogfooding Nightshift by running Claude Code inside a chicklet to migrate production infrastructure, fix a live ETL bug, and ship a full-stack feature. Harness engineering at the platform level.
Containers, But Without The Magic Part 1: Networking
Building container networking from scratch using only the ip command. What network namespaces, veth pairs, and CNI plugins actually do under the hood.
You Are The BIOS Now: Building A Hypervisor In Python With KVM
Build a minimal Type-II hypervisor in Python using KVM. A beginner-friendly walkthrough of ioctls, guest memory, and running your first hello-world VM in about 250 lines.