Deploy Skyvern
AI browser automation with LLMs: API, web UI, Postgres and a login gateway
Just deployed
/var/lib/postgresql/data
skyvern-api
Just deployed
/data
skyvern-ui
Just deployed
Just deployed
Deploy and Host Skyvern on Railway
Skyvern is an open-source tool that automates browser work with LLMs and computer vision. You describe a task in plain language, such as filling a form, downloading invoices or applying to a job, and Skyvern drives a real Chromium browser to do it. You can also chain tasks into workflows and run them on a schedule, from the web UI or the API.
About Hosting Skyvern
Skyvern needs a Python API that runs Chromium inside a virtual display, a web UI, a Postgres database and somewhere to keep screenshots, recordings and downloads. On first start the API runs its database migrations and creates an organization and its API key, and the UI must receive that key before it can serve. The UI has no login of its own, so it must never be exposed directly. This template wires all of that up: the API, UI and database talk over Railway's private network, a small gateway puts a password in front of the UI, and the API's files sit on a volume.
Common Use Cases
- Automate repetitive work on websites that have no API: forms, portals, invoice and report downloads
- Run scheduled browser workflows, such as daily data extraction or monitoring
- Give your own apps or AI agents a browser through the Skyvern API, SDK or MCP server
Dependencies for Skyvern Hosting
- An LLM: an OpenAI, Anthropic or Gemini API key, or any model you add later in Settings (OpenAI-compatible, OpenRouter, Ollama)
- PostgreSQL (included)
Deployment Dependencies
Implementation Details
Services (Skyvern v1.0.58, upstream images pinned by digest, built from the template repo):
- Skyvern: the public gateway (Caddy). It asks for
SKYVERN_USERNAME/SKYVERN_PASSWORDbefore the web UI and passes/api/*calls straight to the API, which checks its own API key. - skyvern-api: the API, the Chromium browser and the task runner. It runs the migrations and creates the organization on first start. Volume
/dataholds artifacts, recordings, downloads, browser sessions and the API key file. - skyvern-ui: the web UI. It waits for the API's first-start key, then serves the app.
- Postgres: the database, on its own volume.
After deploy:
- Wait for
skyvern-apito turn healthy (the first start takes a few minutes). Then open theSkyvernservice's URL and log in with the usernameadminand theSKYVERN_PASSWORDfrom that service's Variables tab. - If you didn't set an LLM key, add a model under Settings, Custom LLMs, or set
OPENAI_API_KEY,ANTHROPIC_API_KEYorGEMINI_API_KEYonskyvern-api. - For the SDK, MCP or REST calls, copy the API key from Settings and call
https://YOUR-DOMAIN/api/v1/...with thex-api-keyheader.
Updates: Railway notifies you when the template repo bumps the Skyvern version. Read CHANGELOG.md in the repo, then apply the update. Migrations run automatically on the next start. Left out on purpose: the optional Vaultwarden/Bitwarden services, the raw VNC port and the Streamlit demo.
Why Deploy Skyvern on Railway?
Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it.
By deploying Skyvern on Railway, you are one step closer to supporting a complete full-stack application with minimal burden. Host your servers, databases, AI agents, and more on Railway.
Template Content
skyvern-api
aalfath/skyvern-railway-templateskyvern-ui
aalfath/skyvern-railway-template
