Station
Self-hosted AI agent runtime that turns 30+ DevOps tools into MCP servers you control.
If your agents touch production infrastructure, Station's shape — runtime on your metal, credentials in declarative variables, 30+ DevOps tools exposed as MCPs under one binary — is the one that survives a security review. The $0 Open Source tier with unlimited agent templates is unusually generous, and CloudShip's own 14-day no-card trial lets you test the dashboard layer before committing. What you're buying with the free tier is control; what you're paying for with your team's time is
Verified 4d ago · liveness 72/100 · cite: rightaichoice.com/tools/station
- DevOps and SRE teams that must keep credentials and infrastructure data inside their own environment
- Platform engineering teams already running Kubernetes, Terraform, and CI/CD who want agents to act on them
- FinOps teams automating AWS cost investigation, anomaly detection, and rightsizing recommendations
- Security-conscious orgs replacing vendor-hosted AI agents with a self-hosted runtime and 30+ DevOps MCPs
- Non-technical users who want a point-and-click agent builder with no YAML or CLI
- Teams unwilling to run and maintain the runtime themselves, including upgrades and uptime
- Buyers shopping for a general-purpose AI assistant or chat app rather than DevOps automation
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Station if you want someone else to host, upgrade, and monitor your agent runtime, or if a YAML-plus-CLI workflow is a dealbreaker for your team.
Centralized dashboards, chat with runs, and cross-context remediation arrive only with the Cloud Platform, which is custom-priced — the $0 tier does not include them.
The $0 Open Source tier fits any team that already runs its own infrastructure and is willing to operate the runtime — unlimited agents with no ceiling. If you need centralized dashboards, cross-context remediation, or leadership reporting, you move to the Cloud Platform at custom pricing, which competes on credential control rather than on seat price against managed agent platforms. Enterprise adds SSO/SAML, an SLA, and on-premise delivery for regulated buyers.
In short
Station — Self-hosted AI agent runtime that turns 30+ DevOps tools into MCP servers you control. Best for DevOps and SRE teams that must keep credentials and infrastructure data inside their own environment, Platform engineering teams already running Kubernetes, Terraform, and CI/CD who want agents to act on them, FinOps teams automating AWS cost investigation, anomaly detection, and rightsizing recommendations. Free to use.
What's new in Station
Checked 4 days agoAcross the latest 4 updates: 4 news mentions.
CloudShip Station: Cursor Security Flaw Reveals AI Agent MCP Risk Pattern
CloudShip argues that the MCP security flaw in Cursor shows black-box AI agents can't be audited, and that teams should self-host AI infrastructure they control.
CloudShip Station: MCP Servers for DevOps Guide 2026
A guide to Model Context Protocol servers for giving AI agents secure access to Kubernetes, Terraform, AWS, and 30+ DevOps tools.
CloudShip Station: 15 Best MCP Servers for DevOps Teams
A roundup of 15 MCP servers for DevOps covering AWS, Kubernetes, GitHub, and monitoring use cases.
CloudShip Station: 7 Reasons Self-Hosted AI Agents Beat SaaS
CloudShip argues self-hosted AI agents beat SaaS for Kubernetes automation and DevOps workflows on security and control.
What people actually say about Station — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
112 mentions across 7 sources (Hacker News, YouTube, Product Hunt, Bluesky, Stack Overflow, Lemmy, Tech Press) · researched Jul 6, 2026.
Average across the 7 sources that answered — each source counts once, not each post.
- +Fully open-source runtime with no vendor lock-in.
- +Self-hosted deployment keeps all credentials and data local.
- +Supports local LLMs, reducing cloud dependency and costs.
- +Pre-built agent templates for 30+ DevOps tools save setup time.
- +Declarative YAML agent builder simplifies configuration.
- −Almost no real user feedback exists to validate claims.
- −Product Hunt users found the tagline and site confusing.
- −Screenshots showed onboarding only, not actual usage.
- −No evidence of an active community or user base.
- −Scaling self-hosted agents may hinder network effects.
- • Cloud platform pricing is undisclosed
- • Self-hosting requires your own infrastructure and maintenance
Viability Score
How well maintained and how widely used is Station? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Self-hosted AI agent runtime that runs on your own infrastructure
- Declarative YAML agent configuration with model, max_steps, tool names, and JSON output schema
- MCP (Model Context Protocol) integration for agent tooling
- 30+ DevOps tools converted into MCP servers under one binary
- Ship CLI installs MCP servers and agent bundles in seconds
- Pre-built agent bundles for FinOps, Security, DevOps, and SRE
- Template registry with ready-to-use agents like aws-cost-explorer
- Declarative variables keep credentials out of templates and AI systems
- Web UI at localhost:8585 for managing tools, bundles, and builds
- Local LLM support via providers such as Ollama, or cloud models like OpenAI
- Chrome extension for browser-based DevOps automation
- Single shell command install and `stn up` startup
- CI/CD gating: agents annotate PRs, fail on high-risk findings, and open autofix PRs
- Policy-as-code bundles using OPA/Conftest, Checkov, and Semgrep
- SBOM generation and license compliance enforcement (Syft, OSV, Grype, ScanCode)
About Station
Station is CloudShip's open-source runtime for AI DevOps agents that execute inside your own infrastructure. DevOps, SRE, platform engineering, and FinOps teams use it to automate runbooks, cost investigations, deployments, and security scans without handing credentials to a third-party AI service. Agents are declared in YAML — the registry's aws-cost-explorer template pins model, max_steps: 15, tool names, and a JSON output schema — then applied with `stn sync environment` and managed from a web UI at localhost:8585. The headline capability is MCP integration. CloudShip has converted 30+ mainstream DevOps tools into MCP servers shipped under one binary and installed with the Ship CLI, covering Terraform, Kubernetes, Docker, Prometheus, Grafana, Vault, ArgoCD, Ansible, Jenkins, Falco, OPA, Istio, Consul, and more. Declarative variables keep secrets out of agent templates, so credentials stay in your environment rather than being passed to a model provider. You can point agents at local LLMs such as Ollama or at cloud providers like OpenAI. Pre-built bundles in the public registry — FinOps, Security, DevOps, and SRE — give you something running on day one. Engineering installs Station once, installs bundles, tailors prompts and thresholds, deploys to their own infrastructure, and monitors runs; leadership then sees structured output through the cloud platform's dashboards and decision views. Station competes less with general-purpose AI assistants and more with managed agent platforms that want your credentials and execution in their cloud. Recent CloudShip posts lean hard on that argument, framing black-box agents as un-auditable and self-hosting as the fix.
Behind the Verdict
We'd reach for Station when the blocker isn't agent capability but agent trust. If your security team has already said no to a SaaS agent holding AWS keys, the pitch here — runtime in your VPC, credentials in declarative variables that never leave your environment, 30+ DevOps tools as MCP servers under one binary — answers the objection directly rather than talking around it. CloudShip's recent blogging doubles down on exactly this: they argue the Cursor MCP flaw proves black-box agents can't be audited, and that self-hosting is the structural fix, not a configuration setting. The day-one experience is genuinely short. Install Station with one curl command, run `stn up --provider openai --api-key`, and you have a web UI at localhost:8585 with bundles for finops, security, devops, and SRE sitting in the registry waiting to install. The aws-cost-explorer template is a good read on how much control you get: model, max_steps, four named tools, and a JSON schema for output, all in YAML you own and can diff in git. Pass on this if nobody on your team wants to own a runtime. Self-hosted means upgrades, uptime, provider keys, and debugging MCP connections are your problem. Managed platforms trade that work away, and for some teams that trade is right. Also pass if your agents aren't DevOps-shaped. This is a runtime for infrastructure work — Terraform plans, Kubernetes ops, SBOM scanning, cost anomalies — not a general assistant. The YAML-and-CLI surface will feel like friction to anyone expecting a point-and-click builder. Where it bites in practice: the control plane question. Agents run on your infrastructure, but the Cloud Platform receives agent run outputs, inputs, and execution metadata. CloudShip is explicit that credentials and sensitive infrastructure data stay
Researching Station? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Station actually fits — and what changes day-one when you adopt it.
Install Station with the Ship CLI, pull the CI guardrails bundle from the public registry, and wire CI Gatekeeper into the PR pipeline so Semgrep and hadolint run on every change.
Outcome: Risky code and Dockerfiles are blocked before merge, and the agent can open autofix PRs instead of filing tickets for a human to triage.
Deploy the aws-cost-explorer agent (model gpt-4o-mini, max_steps: 15) with the AWS cost tools, then run it nightly against Cost Explorer and route structured output to leadership.
Outcome: Spending trends, anomalies, and rightsizing recommendations arrive as structured data rather than a spreadsheet someone rebuilds each month.
Use the dep-lic-sbom bundle to generate CycloneDX SBOMs with Syft, scan with OSV and Grype, and enforce license rules with Conftest on pull requests.
Outcome: Disallowed licenses and known vulnerabilities block the PR, with the SBOM artifact retained as evidence.
Use Cases
- Automate CI/CD pipeline guardrails by running Semgrep SAST, actionlint, and hadolint on every PR, then fail on high-risk findings or open autofix PRs.
- Continuously monitor cloud costs and surface rightsizing and anomaly recommendations from AWS Cost Explorer data.
- Enforce infrastructure-as-code policies using Checkov and OPA/Rego against Terraform plan JSON before apply.
- Detect drift between live cloud resources and IaC, then open PRs to reconcile or import missing resources with evidence tables.
- Generate SBOMs with Syft and enforce license and vulnerability policy with Conftest, OSV, and Grype in the PR pipeline.
- Respond to incidents by automatically gathering logs, analyzing changes, and suggesting fixes.
- Give engineering leadership dashboards and decision views over structured agent run data on the Cloud Platform.
- Run agents against local LLMs such as Ollama so no prompt or output leaves your environment.
Models Under the Hood
as of 2026-10-18
Limitations
- Station requires self-hosting and YAML proficiency; there is no drag-and-drop agent builder.
- The Cloud Platform and Enterprise tiers are custom-priced, so teams that need centralized dashboards, chat with runs, or SSO/SAML must have a scoping conversation with CloudShip rather than check out self-serve.
- The $0 Open Source tier has no centralized management — you manage agents per environment via the localhost:8585 UI and `stn` CLI.
- You own upgrades, uptime, and the operational burden of the runtime; CloudShip does not run it for you outside Enterprise on-premise arrangements.
- Credentials stay local by design, which means the security outcome depends on how well you configure your own environment.
as of 2026-10-04
Verification history
We have re-verified Station 8 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 8 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Station tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Open Source
$0 forever
Cloud Platform
Custom
Ideal for
Teams managing agents across multiple environments that need centralized dashboards, chat with runs, and leadership reporting.
What this tier adds
Adds centralized agent management, chat with agent runs, dynamic dashboards and charts, custom leadership reports, cross-context remediation, multi-agent orchestration, team member management, priority email support, and API access on top of Open Source.
Enterprise
Custom
Ideal for
Regulated organizations that need single sign-on, an uptime commitment, and on-premise or custom-contract delivery.
What this tier adds
Adds SSO and SAML authentication, a dedicated support engineer, custom integrations, a 99.9% uptime SLA, on-premise deployment, custom contract terms, and training and onboarding over Cloud Platform.
Where the pricing makes sense
The company stage and team size where Station's pricing actually pencils out — and where peers do it cheaper.
The $0 Open Source tier fits any team that already runs its own infrastructure and is willing to operate the runtime — unlimited agents with no ceiling. If you need centralized dashboards, cross-context remediation, or leadership reporting, you move to the Cloud Platform at custom pricing, which competes on credential control rather than on seat price against managed agent platforms. Enterprise adds SSO/SAML, an SLA, and on-premise delivery for regulated buyers.
Setup time & first value
How long it actually takes to get something useful out of Station — broken out by persona, not the marketing-page minute.
A platform engineer familiar with Kubernetes and Terraform can install the runtime and apply a first bundle in an afternoon using the Ship CLI and the public registry. First value from a pre-built bundle — CI gatekeeper or SBOM guard — usually lands the same day. Teams that want the Cloud Platform dashboards add a scoping conversation on top of that timeline, and the 14-day trial starts without a
Switching to or from Station
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From a vendor-hosted agent platform: move agent definitions into Station YAML and re-point tools at the equivalent MCP servers from the registry.
- →From hand-rolled scripts and cron jobs: wrap the existing runbook as a Station agent with a model, max_steps, and output schema.
- →From a generic AI assistant used for DevOps questions: replace prompt-driven answers with a Station agent that has real tool access to your cluster and cloud account.
- →From CI plugins for SAST and linting: consolidate into the ci-guardrails bundle so one agent runs Semgrep, actionlint, hadolint, and Trivy together.
- ↗To a managed agent platform: export agent YAML as documentation of your intended policy, then re-implement guardrails in the vendor's rule engine.
- ↗To self-managed CI policy scripts: swap the registry bundles for your own Semgrep and Checkov invocations and keep the CI gating logic.
- ↗To a general-purpose AI assistant: keep the agents as reference runbooks but accept that you lose tool-level execution against your infrastructure.
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Station”, and we withheld 6: 6 could not be judged, because “Station” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Station.
Official links
Featured Head-to-Head Comparisons
Station vs Spider Cloud
Choose Spider Cloud if your primary need is high-volume web scraping for AI agents or RAG pipelines — its Rust engine, Browser AI commands, and pay-as-you-go model deliver speed and flexibility. Choose Station if you’re a DevOps team wanting to run AI agents fully on your own infrastructure with MCP integration to tools like Kubernetes and Terraform, and you can handle YAML-based configuration.
Station vs Temporal Ai
Temporal AI wins for teams needing bulletproof workflow reliability for AI agents and microservices, with automatic retries and state persistence. Station is ideal for DevOps teams that want self-hosted, no-outbound-data AI agent automation on their infrastructure. Choose Temporal if uptime and fault tolerance matter most; pick Station if you demand full control over agent data and infrastructure.
Station vs Presto Voice
Presto Voice and Station serve completely different domains. Presto Voice is a specialized drive-thru voice AI for QSR chains, proven to boost revenue via upselling and high automation rates. Station is an open-source DevOps agent runtime for teams that want self-hosted, secure automation with full control over data and infrastructure. Choose Presto if you run a multi-location QSR and want measurable sales lift; choose Station if you’re a DevOps team wanting customizable, on-premise AI agents.
Popular in AIOps & Incident Response
Dash0
Dash0 is OpenTelemetry-native observability where Agent0 investigates incidents, finds root cause, and opens fix PRs.
Resolve AI
AI SRE platform whose autonomous agents go on-call, triage alerts, investigate incidents and run production tasks alongside your engineers.
Frequently Asked Questions
Used Station? Help shape our editorial sentiment research.