--- title: "Deploy Langfuse v3.226.0 (Production ready AI Observability)" description: "Open-source LLM platform for prompt management, tracing, evals & metrics" category: "AI/ML" url: https://railway.com/deploy/kUsf_w --- # Deploy Langfuse v3.226.0 (Production ready AI Observability) Open-source LLM platform for prompt management, tracing, evals & metrics **[Deploy Langfuse v3.226.0 (Production ready AI Observability) on Railway](https://railway.com/template/kUsf_w)** Machine-readable deploy manifest (JSON, validated by TemplateCI): https://railway.com/deploy/kUsf_w/manifest.json - **Creator:** RailGenius - **Category:** AI/ML - **Total deploys:** 44 ## Template content ### clickhouse https://devicons.railway.app/ClickHouse?variant=light - **Image:** clickhouse/clickhouse-server:25.11 - **Health check:** /ping ### postgres https://devicons.railway.app/i/postgresql.svg - **Image:** ghcr.io/railwayapp-templates/postgres-ssl:16.11 ### redis https://devicons.railway.app/i/redis.svg - **Image:** bitnami/redis:latest - **Start command:** `/bin/sh -c 'mkdir -p /bitnami/redis/data && exec /opt/bitnami/redis/bin/redis-server --bind :: 0.0.0.0 --requirepass "$REDIS_PASSWORD" --dir /bitnami/redis/data --appendonly yes'` ### langfuse-web https://langfuse.com/icon.svg - **Image:** langfuse/langfuse:3.226.0 - **Health check:** /api/public/health - **Public domain:** Yes ### minio https://res.cloudinary.com/h5kvzw/image/upload/v1706062979/MINIO_Bird_vbtpet.png - **Image:** chrislusf/seaweedfs@sha256:f7a49e3f3972ebed5be2ed4befa4a0c68eb7bdddba6e692d42729ceb545134b8 - **Start command:** `/usr/bin/weed mini -dir=/data -ip=127.0.0.1 -ip.bind=0.0.0.0 -s3.port=9000 -master.telemetry=false -s3.port.iceberg=0 -s3.port.lance=0` - **Public domain:** Yes ### langfuse-worker https://langfuse.com/icon.svg - **Image:** langfuse/langfuse-worker:3.226.0 ## Documentation # Deploy and Host Langfuse v3.226.0 on Railway Self-host Langfuse for LLM tracing and observability, with application and worker images aligned to v3.226.0. ## About Hosting Langfuse v3.226.0 This template includes Langfuse Web, Langfuse Worker, PostgreSQL, ClickHouse, authenticated Redis and persistent SeaweedFS S3 storage. The storage service retains the legacy name `minio` to preserve template references. Wait for initial database migrations and the `/api/public/health` readiness check, then open the generated web domain and create your account and project. Encryption keys, storage credentials, database passwords and session secrets are generated per installation. Web and worker use consistent encryption/salt references. This listing remains on the v3 line and does not migrate to v4. ## First deployment: required steps 1. Deploy and wait for the databases, storage, web and worker services to initialize. Database migrations can take several minutes. 2. Open the Railway-generated HTTPS domain on `langfuse-web`, then create your user, organization and project promptly. 3. Create project API keys in Langfuse and point your application's integration at this self-hosted web URL. Keep the secret key server-side. 4. Send a sample trace and confirm that it appears in your project before sending production traffic. This listing stays on Langfuse v3. Use a v3-compatible SDK/integration and this deployment’s project API keys. Upgrading to v4 is a separate migration. ### Already configured for a new installation You do not need to create the S3 bucket manually, run a MinIO client, generate infrastructure passwords or encryption keys, wire database connections, or add custom domains. These are supplied by the template. The service named `minio` runs SeaweedFS; it has no MinIO console or MinIO bucket-setup step. Keep the generated keys and service references intact. Custom domains, email delivery, external providers, retention and backups are optional configuration or operational decisions. If you change the web domain, update `NEXTAUTH_URL` and your clients to the new HTTPS origin. If you change storage, update both web and worker S3 references consistently; retain a browser-accessible signed-media endpoint. ## Why Deploy Langfuse v3.226.0 on Railway Provision the connected services and generated credentials in one deployment, with persistent storage and private service references configured for you. RailGenius maintains this community template; it is not an official upstream-hosted service. Clean installations passed signup/login, authenticated trace ingestion/querying, signed media upload/download and unsigned-object rejection. Stored objects survived replacement of the storage container. Optional external model providers and additional evaluation runners were not exercised. ## Common Use Cases - Collect and query LLM application traces. - Inspect inputs, outputs and trace metadata. - Upload and retrieve multimodal trace attachments with signed URLs. ## Dependencies for Langfuse v3.226.0 Event uploads use private S3 networking. Media uploads/downloads use a public HTTPS S3 endpoint with authenticated or presigned access. Storage administration ports are not publicly routed. Redis requires its generated password and persists append-only data on its volume. ### Deployment Dependencies - A Railway account with sufficient resources for the application, databases, workers and volumes. - Generated Railway domains for the public endpoints; no custom domain is required for initial setup. - Your own credentials for optional third-party integrations or delivery providers. - Upstream project documentation and license: [Langfuse source](https://github.com/langfuse/langfuse) and [self-hosting documentation](https://langfuse.com/self-hosting). ## Storage, backups and existing installations New installations use persistent SeaweedFS S3 storage instead of unavailable MinIO images. Keep its complete `/data` volume, database backups and application encryption/session keys in your recovery plan. This template update applies to new installations. Existing projects are not automatically upgraded. MinIO files cannot be mounted directly into SeaweedFS: existing deployments need a separate authenticated S3 export/copy, verification and rollback plan. Preserve existing encryption keys when retaining a database. ## Costs and retention Usage depends on application/worker activity, database memory and CPU, persistent volumes, uploads and retained data. Configure retention and backups for your workload. Your deployed project starts accruing Railway usage even while idle; the listing itself does not include a running demo. Storage repair and clean-install validation: 1 November 2026. Compatible v3.226.0 release validation: 3 November 2026. Application and worker remain on the supported v3 line; this does not migrate an existing installation to v4. ## Similar templates - [Chat Chat](https://railway.com/deploy/-WWW5r) — Chat Chat, your own unified chat and search to AI platform. - [Weaviate](https://railway.com/deploy/weaviate-1) — Vector database with hybrid search and an API key-protected API - [Möbius v0.1 Personal AI OS](https://railway.com/deploy/mobius-v01-personal-ai-os) — Self-hosted AI agent that builds and runs apps for you, with memory. Open this page in a browser: https://railway.com/deploy/kUsf_w