Langflow on Railway
Visual builder for AI agents, RAG pipelines and MCP servers, running Langflow 1.12.5 on Railway.
Langflow is an open-source visual builder for AI agents, retrieval-augmented generation (RAG) pipelines and MCP servers. You drag components onto a canvas, connect models, tools and data sources, test the flow in a playground and expose it through a REST API or as an MCP server.
It is used to prototype LLM agents without writing a backend, to put chat over your own documents, and to give a team a shared workspace for AI workflows. It is often mentioned alongside LangChain, whose components it builds on, but it is a separate visual product.
Langflow is released under the MIT licence.
What Langflow can do
- Visual canvas for agents and RAG flows
- REST API for every flow
- Publish flows as MCP servers
- Global variables encrypted at rest
- Embeddable chat widget
- Multi-worker mode backed by Redis
What people use it for
- Prototyping LLM agents with tools such as web search and APIs
- Building retrieval-augmented chat over your documents and exposing it as an API
- Publishing flows as MCP servers for Claude Desktop, Cursor and other clients
- A shared, login-protected workspace for AI workflows
Videos
Running Langflow on Railway
The Railway template runs langflowai/langflow 1.12.5 with Postgres 18 for flows, users, messages and encrypted variables instead of SQLite, Redis 8.2 as the shared build and job queue, and a volume at /app/langflow for uploaded files, knowledge bases and logs. It starts with two Gunicorn workers and a generated admin password and encryption key. Memory grows with the number of workers, at roughly 0.6 to 1 GB each according to the template notes.
Langflow templates on Railway
Tracklayer templates are listed first. Other Langflow templates follow, ordered by total deployments on Railway. How this works.
Deployment numbers as of 2026-11-20.
Questions about Langflow
How do I log in?
Open the public URL and sign in as admin with LANGFLOW_SUPERUSER_PASSWORD from the Langflow service variables. The first deploy can take a few minutes while migrations and starter projects load.
Where do I put model provider keys?
Under Settings, Global Variables. They are encrypted with LANGFLOW_SECRET_KEY, which you must never change after the first boot.
How do I add more workers?
Raise LANGFLOW_WORKERS, about one per vCPU, and give the service more resources. Keep LANGFLOW_WORKERS times the pool size and max overflow below about 90 database connections.
Can I embed the chat widget on my site?
Yes. Add each website origin that embeds Langflow to LANGFLOW_CORS_ORIGINS. For a custom domain, also update LANGFLOW_MCP_BASE_URL.
How do I upgrade?
Back up Postgres, change the image tag on the Langflow service and redeploy. Langflow migrates the database on start, so do not roll back to an older version against a migrated database without restoring the backup.
Guides
Similar apps
Tracklayer maintains the templates marked as such; Langflow is a trademark of its owners and this site is not affiliated with them.