AI tools & infrastructure on Railway

AI applications, model operations, inference and supporting vector databases.

7 documented options on Railway. Provider coverage reflects reviewed evidence, not an exhaustive compatibility claim.

ListingDocumented cloudsDeployment and infrastructureLicensingInstallationOperations and limitations
AnythingLLM DockerAWS, Google Cloud, RailwayCustomer-operated
Servers / containers
Open Source / Open Core

MIT License for upstream AnythingLLM; model weights, connectors and provider services have separate terms.

Run the official Docker edition on a customer-operated host with persistent application storage and configured model/vector endpoints.Customer manages users, documents, persistent storage, model access, integration secrets, backups and upgrades.
Scope and evidence

The Docker edition supports multi-user operation and differs from the desktop application. Local inference is an option; configured external LLMs, embeddings, vector stores and agent tools can send data outside the account. AnythingLLM v1.20.0 hardens embed history, extension API-key authorization, workspace-agent access, password recovery, framing, and document-tool downloads. Optional MCP identity forwarding creates a new outbound identity path; keep the server and extension on compatible 1.20+ versions.

  • AnythingLLM Docker installation overview — The project explicitly names AWS, Google Cloud and Railway for self-hosting the Docker edition, paired with its Docker installation workflow. This is an upstream hosting compatibility statement. Model endpoints, embeddings and external vector stores are separately configured and can have different data boundaries.

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

Cloud evidence reviewed 2026-11-12

Arize Phoenix Self-HostedAWS, Azure, Google Cloud, Railway, RenderCustomer-operated
Servers / containers
Source Available

Elastic License 2.0; includes restrictions on providing the software as a hosted/managed service and on license-key circumvention. This is source-available.

Use the official self-hosted Docker/Compose or Kubernetes deployment instructions and configure persistent storage and instrumentation endpoints.Customer manages trace capture, storage, authentication, retention, evaluation providers, backups and upgrades.
Scope and evidence

The self-hosted product is distinct from commercial Arize offerings. Traces may contain prompts and outputs; evaluation jobs can call separately configured external model services even when trace storage is local.

  • Phoenix self-hosting deployment options — The current Phoenix self-hosting index includes a Railway one-click template as well as AWS CloudFormation, Azure ARM and Google Cloud Run. These deploy the Phoenix application, not the separately licensed Arize AX product. A Railway project is customer-controlled while Railway manages its infrastructure.

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

  • Phoenix self-hosted Render Blueprint — The vendor self-hosting index offers a Render one-click Blueprint for Phoenix alongside its Railway, AWS, Azure and Google Cloud options. This is the self-hosted Phoenix application, separate from Arize AX; customer workspace control does not imply ownership of the platform infrastructure.

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

Cloud evidence reviewed 2026-11-12

Langfuse Self-HostedAWS, Azure, Google Cloud, Railway, RenderCustomer-operated
Servers / containers
Open Source / Open Core

MIT core; designated enterprise directories and add-on capabilities use separate license terms or keys.

Docker Compose for a single VM; Helm/Kubernetes and official AWS, Azure and Google Cloud Terraform modules.Customer operates the deployment, persistence, access policies, retention, backups and scaling.
Scope and evidence

Traces can contain prompts, outputs and other sensitive application data, so configure capture and retention deliberately. Self-hosted trace storage does not move the underlying application model into the same account. Optional model-based evaluators or the gateway use configured providers, which may be external. The single-VM Compose example does not provide production high availability or backups. Langfuse v4.82.0 fixes an SSRF path in GitHub Enterprise SSO and moves audit logs to ClickHouse. Review outbound SSO connectivity and audit-log storage during upgrade.

  • Langfuse customer cloud deployment on AWS — The vendor documents deployment of the self-hosted Langfuse application and its backing services in the customer account or project, including provider-specific infrastructure. This is distinct from Langfuse Cloud; database, object storage, model calls and enterprise licensing must be configured for the chosen topology.

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

  • Langfuse customer cloud deployment on Azure — The vendor documents deployment of the self-hosted Langfuse application and its backing services in the customer account or project, including provider-specific infrastructure. This is distinct from Langfuse Cloud; database, object storage, model calls and enterprise licensing must be configured for the chosen topology.

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

  • Langfuse customer cloud deployment on Google Cloud — The vendor documents deployment of the self-hosted Langfuse application and its backing services in the customer account or project, including provider-specific infrastructure. This is distinct from Langfuse Cloud; database, object storage, model calls and enterprise licensing must be configured for the chosen topology.

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

  • Langfuse v3 community Railway deployment — The vendor documents a community-maintained Railway template for the complete Langfuse v3 stack. This establishes that version-specific application deployment; the same page explicitly distinguishes the v4 alternatives. The vendor does not systematically test the community template. Review the template services, image tags, database timezones and upgrade path.

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

  • Langfuse v4 community Render Blueprint — The vendor documents a community-maintained v4 Blueprint with separate web and worker services, PostgreSQL, Key Value, ClickHouse and MinIO. It creates the full application in the customer Render workspace. Review service plans, persistent disks, object-storage exposure and recovery procedures; community support is best-effort.

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

Cloud evidence reviewed 2026-11-12

LibreChat Self-HostedRailway, DigitalOceanCustomer-operated
Servers / containers
Open Source / Open Core

MIT

Official Railway one-click template or Docker on a DigitalOcean Ubuntu DropletCustomer manages the Railway project or DigitalOcean Droplet, application secrets, model credentials, data services, backups, scaling and upgrades. Railway operates its platform; Droplet users also maintain the operating system and Docker runtime.
Scope and evidence

The Railway path runs in the customer's project on platform-managed infrastructure; the DigitalOcean path uses a customer-operated Droplet. External model, embedding, search and tool providers receive data according to customer configuration. Review generated secrets, public networking, persistent data services and image updates before production use.

  • LibreChat one-click deployment on Railway — LibreChat documents creating a Railway account, reviewing the template's three containers and environment variables, deploying the services and accessing the resulting LibreChat URL in the customer's Railway project.

    Vendor / upstream documentation · Reviewed 2026-11-15. Vendor support commitment: not assessed; confirm the selected release and contract.

  • LibreChat deployment on DigitalOcean — The upstream guide creates an Ubuntu Droplet in the customer DigitalOcean project, configures a user and firewall, then links to the Docker installation steps. The example is for a small personal deployment, not a production sizing guarantee.

    Vendor / upstream documentation · Reviewed 2026-11-15. Vendor support commitment: not assessed; confirm the selected release and contract.

Cloud evidence reviewed 2026-11-15

LiteLLM Proxy Self-HostedAWS, Azure, Google Cloud, Railway, RenderCustomer-operated
Servers / containers
Open Source / Open Core

MIT outside the enterprise directory; enterprise code and features have separate license terms.

Official container image, Docker Compose or Kubernetes deployment; production infrastructure examples are documented.Customer manages gateway availability, credentials, database, access policies, logs and upgrades, including pinning versions and verifying release artifacts with the documented version-appropriate Cosign method; model providers remain separate services.
Scope and evidence

The gateway forwards each request to its configured model provider. Self-hosting the proxy does not ensure prompts stay inside the account: external model APIs, guardrail services and logging integrations can receive data. Use appropriate self-hosted inference endpoints and explicitly review outbound integrations when required. For release integrity, release images and the Helm chart starting with v1.111.0-dev.3 use keyless Cosign signatures bound to LiteLLM's release workflow; earlier releases require the documented historical public keys. The October 21 incident affected PyPI v1.107.0 and Docker Hub/GHCR v1.107.0 plus mutable latest, main-latest and main-stable tags during the documented exposure windows. Check the incident's artifact digests and remediation instructions; the compromised images carried valid historical signatures, so signature verification alone does not establish safety. LiteLLM v1.110.0 changes authentication, guardrail, accounting, and S3 behavior. The 1.108.11 LTS patch also fixes Bedrock guarded_text handling and restricts os.environ team settings to proxy administrators; review both the selected release line and existing incident-response or image-signing guidance. LiteLLM v1.108.12 LTS fixes tool_policy allowing blocked /v1/messages tool names longer than 64 characters. The standard v1.110.1 UI image changes to slim nginx and removes curl while retaining wget and nc, so retest health probes and scripts that assume curl.

  • LiteLLM production deployment — Official production guidance covers EKS, AKS and GKE with Helm, plus Terraform on AWS and Google Cloud. The proxy, database and Redis deployment is separate from the model services it calls.

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

  • LiteLLM Railway and Render deployment options — The current upstream deployment guide supplies Railway and Render deployment buttons in addition to its EKS/AKS/GKE instructions, followed by readiness and Admin UI verification. These are customer application deployments on managed platforms; configure persistent data, secrets and external model-provider endpoints separately.

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

Cloud evidence reviewed 2026-11-12

OpenSearchRailwayCustomer-operated
Servers / containers
Open Source / Open Core

Apache License 2.0 for upstream OpenSearch. Separately operated managed services and connected model or data services have their own terms.

Run official Docker images and customize the documented Docker Compose examples; native package and tarball installations are also documented.Customer manages cluster sizing, storage, shard and replica layout, TLS, users and roles, ingestion, snapshots, monitoring and upgrades.
Scope and evidence

Sample deployments and demo security settings require production configuration. Official images distributed through Amazon ECR are not evidence of an AWS-specific deployment. The listing covers self-operated OpenSearch; Amazon OpenSearch Service and other managed services are separate delivery models. Embeddings and AI connectors can involve external models or services. OpenSearch 3.10.0 removes experimental Views and macOS x64 archives and fixes remote-store data-loss or stall paths, bulk-index corruption, heap leaks, snapshot queue stalls, remote-translog issues, and OOM behavior. Validate plugins, Dashboards, and storage recovery before upgrading.

  • Self-operated OpenSearch and Dashboards on Railway — The community recipe deploys OpenSearch 3 and Dashboards as customer-project services, with the search API on private networking. Its wrappers derive from upstream OpenSearch images. Review node count, heap/storage sizing, security plugin configuration, TLS and backups; this is separate from Amazon OpenSearch Service and does not imply high availability.

    Community template · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

Cloud evidence reviewed 2026-11-12

RAGFlowAWS, Azure, Google Cloud, Alibaba Cloud, RailwayCustomer-operated
Servers / containers
Open Source / Open Core

Apache License 2.0 for upstream RAGFlow; models, data assets and dependencies retain their own terms.

Versioned upstream Docker Compose, provider-specific VM packages or community deployment templates; select a path compatible with the intended RAGFlow release.Customer manages document storage, indexing, app services, access controls, model endpoints, backups and upgrades.
Scope and evidence

Deployment options include customer AWS, Azure and Google Cloud infrastructure, Alibaba Compute Nest, and a Railway project. The evidence trail identifies the publishers and template versions. Hosting RAGFlow does not automatically self-host its LLM, embeddings or reranker; choose private or local model endpoints where required.

  • Community RAGFlow deployment on AWS ECS Fargate through stdapi.ai — The sample provisions the RAGFlow application and workers, OpenSearch, Aurora PostgreSQL, Valkey, S3 and a Bedrock gateway in the customer AWS account. Source inspection confirms the v0.26.4 Python image pin; it does not establish compatibility with the 1.0 Go rewrite. This community integration also requires the separately billed stdapi.ai gateway and Bedrock access. Review its documented networking, client-security, availability and cost limitations before use.

    Community template · Reviewed 2026-11-12 · Scope: Community / upstream open-source; third-party distribution where noted · Version: 0.26.4. Vendor support commitment: not assessed; confirm the selected release and contract.

  • cloudimg RAGFlow Azure VM deployment and operations guide — The image publisher documents an Azure VM in the customer subscription running the complete Compose stack, including Elasticsearch, MySQL, MinIO, Redis and a local embedding service. Its guide pins RAGFlow v0.26.4, not 1.0. This is a third-party repackaged image with additional support charges, not an upstream-endorsed distribution. Confirm image access, current patch/migration requirements, TLS configuration and the separately configured chat-model endpoint.

    Third party package · Reviewed 2026-11-12 · Scope: Community / upstream open-source; third-party distribution where noted · Version: 0.26.4. Vendor support commitment: not assessed; confirm the selected release and contract.

  • Techlatest RAGFlow VM deployment in a customer Google Cloud project — The image publisher documents selecting the customer project, deployment service account, zone, network and Compute Engine VM, then completing RAGFlow initialization and configuring its bundled Ollama models. This is a third-party packaged deployment; the guide does not establish a specific current RAGFlow release or 1.0 compatibility. Confirm the image version, support charges, model licenses, resource sizing and network exposure before use.

    Third party package · Reviewed 2026-11-12 · Scope: Community / upstream open-source; third-party distribution where noted. Vendor support commitment: not assessed; confirm the selected release and contract.

  • RAGFlow Community deployment on Alibaba Cloud Compute Nest — Alibaba Cloud documents provisioning a RAGFlow Community service instance with customer ECS, VPC and vSwitch parameters. This is distinct from the managed AI-Native Database Service integration. Review image version, access controls, model endpoints and the broad example RAM permissions before production.

    Provider guide · Reviewed 2026-11-12 · Scope: Community / upstream open-source; third-party distribution where noted. Vendor support commitment: not assessed; confirm the selected release and contract.

  • Community RAGFlow Railway application stack — This community template deploys the application, Elasticsearch, MySQL, Redis, a local embedding service and object storage into the customer Railway project. Its linked Dockerfile pins upstream RAGFlow v0.27.1 with a Railway startup/configuration wrapper. The template is not proof of 1.0 compatibility or upstream support; review migrations and the separately configured chat-model endpoint. Railway manages the underlying infrastructure.

    Community template · Reviewed 2026-11-12 · Scope: Community / upstream open-source; third-party distribution where noted · Version: 0.27.1. Vendor support commitment: not assessed; confirm the selected release and contract.

  • Railway deployment image source and release pin — Source inspection confirms the v0.27.1 base image and the custom configuration, nginx and startup wrapper used by the linked Railway template. The upstream 1.0 release changes the architecture and requires its documented migration path; do not substitute a new image tag without reviewing the complete stack.

    Community template · Reviewed 2026-11-12 · Scope: Community / upstream open-source; third-party distribution where noted · Version: 0.27.1. Vendor support commitment: not assessed; confirm the selected release and contract.

Cloud evidence reviewed 2026-11-12

Deployment evidence and category fit

AnythingLLM Docker

Self-hosted AI application interface and retrieval/agent integrations; model inference is separately configured.

The Docker edition supports multi-user operation and differs from the desktop application. Local inference is an option; configured external LLMs, embeddings, vector stores and agent tools can send data outside the account. AnythingLLM v1.20.0 hardens embed history, extension API-key authorization, workspace-agent access, password recovery, framing, and document-tool downloads. Optional MCP identity forwarding creates a new outbound identity path; keep the server and extension on compatible 1.20+ versions.

  • AnythingLLM Docker installation overview ↗

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

    The project explicitly names AWS, Google Cloud and Railway for self-hosting the Docker edition, paired with its Docker installation workflow. This is an upstream hosting compatibility statement. Model endpoints, embeddings and external vector stores are separately configured and can have different data boundaries.

Cloud evidence reviewed 2026-11-12. Full listing and qualifications

Arize Phoenix Self-Hosted

AI development and operations tooling for tracing and evaluation rather than a model-serving runtime.

The self-hosted product is distinct from commercial Arize offerings. Traces may contain prompts and outputs; evaluation jobs can call separately configured external model services even when trace storage is local.

  • Phoenix self-hosting deployment options ↗

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

    The current Phoenix self-hosting index includes a Railway one-click template as well as AWS CloudFormation, Azure ARM and Google Cloud Run. These deploy the Phoenix application, not the separately licensed Arize AX product. A Railway project is customer-controlled while Railway manages its infrastructure.

Cloud evidence reviewed 2026-11-12. Full listing and qualifications

Langfuse Self-Hosted

Self-hosted LLM observability, evaluations and prompt-management platform.

Traces can contain prompts, outputs and other sensitive application data, so configure capture and retention deliberately. Self-hosted trace storage does not move the underlying application model into the same account. Optional model-based evaluators or the gateway use configured providers, which may be external. The single-VM Compose example does not provide production high availability or backups. Langfuse v4.82.0 fixes an SSRF path in GitHub Enterprise SSO and moves audit logs to ClickHouse. Review outbound SSO connectivity and audit-log storage during upgrade.

  • Langfuse v3 community Railway deployment ↗

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

    The vendor documents a community-maintained Railway template for the complete Langfuse v3 stack. This establishes that version-specific application deployment; the same page explicitly distinguishes the v4 alternatives. The vendor does not systematically test the community template. Review the template services, image tags, database timezones and upgrade path.

Cloud evidence reviewed 2026-11-12. Full listing and qualifications

LibreChat Self-Hosted

Provides a self-hosted interface and orchestration layer for model endpoints, agents, MCP tools, authentication and administrative controls.

The Railway path runs in the customer's project on platform-managed infrastructure; the DigitalOcean path uses a customer-operated Droplet. External model, embedding, search and tool providers receive data according to customer configuration. Review generated secrets, public networking, persistent data services and image updates before production use.

  • LibreChat one-click deployment on Railway ↗

    Vendor / upstream documentation · Reviewed 2026-11-15. Vendor support commitment: not assessed; confirm the selected release and contract.

    LibreChat documents creating a Railway account, reviewing the template's three containers and environment variables, deploying the services and accessing the resulting LibreChat URL in the customer's Railway project.

Cloud evidence reviewed 2026-11-15. Full listing and qualifications

LiteLLM Proxy Self-Hosted

Self-hosted AI API gateway providing model routing, access controls and usage tracking.

The gateway forwards each request to its configured model provider. Self-hosting the proxy does not ensure prompts stay inside the account: external model APIs, guardrail services and logging integrations can receive data. Use appropriate self-hosted inference endpoints and explicitly review outbound integrations when required. For release integrity, release images and the Helm chart starting with v1.111.0-dev.3 use keyless Cosign signatures bound to LiteLLM's release workflow; earlier releases require the documented historical public keys. The October 21 incident affected PyPI v1.107.0 and Docker Hub/GHCR v1.107.0 plus mutable latest, main-latest and main-stable tags during the documented exposure windows. Check the incident's artifact digests and remediation instructions; the compromised images carried valid historical signatures, so signature verification alone does not establish safety. LiteLLM v1.110.0 changes authentication, guardrail, accounting, and S3 behavior. The 1.108.11 LTS patch also fixes Bedrock guarded_text handling and restricts os.environ team settings to proxy administrators; review both the selected release line and existing incident-response or image-signing guidance. LiteLLM v1.108.12 LTS fixes tool_policy allowing blocked /v1/messages tool names longer than 64 characters. The standard v1.110.1 UI image changes to slim nginx and removes curl while retaining wget and nc, so retest health probes and scripts that assume curl.

  • LiteLLM Railway and Render deployment options ↗

    Vendor / upstream documentation · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

    The current upstream deployment guide supplies Railway and Render deployment buttons in addition to its EKS/AKS/GKE instructions, followed by readiness and Admin UI verification. These are customer application deployments on managed platforms; configure persistent data, secrets and external model-provider endpoints separately.

Cloud evidence reviewed 2026-11-12. Full listing and qualifications

OpenSearch

Vector storage, semantic search and retrieval for AI applications and RAG. Embedding generation and model connectors have separately configured runtimes and data flows.

Sample deployments and demo security settings require production configuration. Official images distributed through Amazon ECR are not evidence of an AWS-specific deployment. The listing covers self-operated OpenSearch; Amazon OpenSearch Service and other managed services are separate delivery models. Embeddings and AI connectors can involve external models or services. OpenSearch 3.10.0 removes experimental Views and macOS x64 archives and fixes remote-store data-loss or stall paths, bulk-index corruption, heap leaks, snapshot queue stalls, remote-translog issues, and OOM behavior. Validate plugins, Dashboards, and storage recovery before upgrading.

  • Self-operated OpenSearch and Dashboards on Railway ↗

    Community template · Reviewed 2026-11-12. Vendor support commitment: not assessed; confirm the selected release and contract.

    The community recipe deploys OpenSearch 3 and Dashboards as customer-project services, with the search API on private networking. Its wrappers derive from upstream OpenSearch images. Review node count, heap/storage sizing, security plugin configuration, TLS and backups; this is separate from Amazon OpenSearch Service and does not imply high availability.

Cloud evidence reviewed 2026-11-12. Full listing and qualifications

RAGFlow

Self-hosted document ingestion, retrieval and context infrastructure for AI applications.

Deployment options include customer AWS, Azure and Google Cloud infrastructure, Alibaba Compute Nest, and a Railway project. The evidence trail identifies the publishers and template versions. Hosting RAGFlow does not automatically self-host its LLM, embeddings or reranker; choose private or local model endpoints where required.

  • Community RAGFlow Railway application stack ↗

    Community template · Reviewed 2026-11-12 · Scope: Community / upstream open-source; third-party distribution where noted · Version: 0.27.1. Vendor support commitment: not assessed; confirm the selected release and contract.

    This community template deploys the application, Elasticsearch, MySQL, Redis, a local embedding service and object storage into the customer Railway project. Its linked Dockerfile pins upstream RAGFlow v0.27.1 with a Railway startup/configuration wrapper. The template is not proof of 1.0 compatibility or upstream support; review migrations and the separately configured chat-model endpoint. Railway manages the underlying infrastructure.

  • Railway deployment image source and release pin ↗

    Community template · Reviewed 2026-11-12 · Scope: Community / upstream open-source; third-party distribution where noted · Version: 0.27.1. Vendor support commitment: not assessed; confirm the selected release and contract.

    Source inspection confirms the v0.27.1 base image and the custom configuration, nginx and startup wrapper used by the linked Railway template. The upstream 1.0 release changes the architecture and requires its documented migration path; do not substitute a new image tag without reviewing the complete stack.

Cloud evidence reviewed 2026-11-12. Full listing and qualifications

Deployment and procurement checks

Project networking
Private networking connects services inside a project environment using encrypted tunnels and internal DNS. This project boundary does not give the customer ownership of the underlying infrastructure account. Provider source ↗ · Reviewed 2026-11-14

For every provider, confirm marketplace/private-offer terms, commitment eligibility, regional service availability, private connectivity and any vendor access against your selected deployment. Provider capabilities do not automatically establish application support.

Plan this deployment

Account boundary: Railway workspace, project and environment; platform-managed infrastructure. Access: Workspace/project access and scoped deployment tokens. Network: Project/environment private networking with separately exposed public services.

Provider documentation ↗ · Guidance reviewed 2026-11-12. Confirm the application edition, region, support access, backup locations and operating responsibilities before choosing.

Compare this category across clouds · Search these listings · Compare 2–4 listings