LangSmith BYOC Expands to Azure and Google Cloud for AI Agents

The infrastructure overhead of running autonomous AI agents just hit a speed bump for multi-cloud teams. According to the LangChain Blog, LangSmith Bring Your Own Cloud (BYOC) has expanded its reach, moving from an AWS-only release to general availability on AWS and beta availability on both Azure and Google Cloud.

What happened

LangChain announced that its BYOC model—which splits platform management between a LangChain-run control plane and a user-owned data plane—now spans all three major cloud providers. According to the LangChain Blog, the control plane handles authentication, organization and workspace configuration, billing metadata, hashed API keys, and static UI assets. Meanwhile, the data plane resides entirely within the customer's cloud account and virtual network, securing sensitive assets like traces, prompts, datasets, experiments, agent deployments, and workspace secrets.

The expansion brings specific native implementations for each provider. According to the company, Azure BYOC runs inside a dedicated virtual network with a private AKS API endpoint, utilizing Istio and KEDA with Helm for ingress routing and event-driven scaling, alongside Azure Monitor for system telemetry. Google Cloud data planes can run in any region using namespace-scoped names to prevent collisions. AWS deployments run directly inside existing VPCs with databases in user-selected subnets, allowing security teams to attach custom security groups.

Previously, small teams with strict compliance needs had to deploy and maintain self-hosted LangSmith instances, requiring dedicated platform engineers to patch stateful services and run schema migrations. With BYOC, LangChain provisions and operates the infrastructure inside the customer's virtual network while handling scaling, schema updates, and runtime dependencies. Organizations can migrate existing self-hosted data using an open-source migration tool and internal tooling provided by LangChain.

Why it matters for small engineering teams

For a small team building agentic workflows, compliance and infrastructure maintenance have always been at war. You either spent precious engineering hours nursing Kubernetes clusters and Helm charts for self-hosted tools, or you compromised by sending sensitive agent traces and prompts to a SaaS vendor's multi-tenant cloud.

BYOC finally bridges this gap without requiring a platform engineering department. By letting LangChain manage the data plane's operational lifecycle—scaling, patches, and updates—while keeping traces and prompts locked inside your own Azure, AWS, or GCP account, you get enterprise-grade compliance with zero operational drag. If your small team is deploying agents that touch proprietary business data, this architecture lets you satisfy security stakeholders in minutes instead of dedicating a full-time dev to infrastructure maintenance.

What to do this week

1. Audit your current agent telemetry and storage boundaries to determine which cloud provider your team's production and staging workloads currently inhabit.
2. Review the public Terraform modules available in LangChain's GitHub repository to estimate the implementation effort for deploying a data plane in your chosen cloud environment.
3. Contact LangChain sales to discuss provisioning a BYOC data plane and evaluate their open-source migration tool if you are currently running a self-hosted instance.

Sources

• LangSmith BYOC is now available on AWS, Azure, and Google Cloud — LangChain Blog