LangChain Opens Tool Gateway Beta to Police AI Agent Credentials
LangChain's LangSmith Tool Gateway entered public beta on November 19, 2026, centralizing agent authentication, tool access rules and audit trails for enterprise AI deployments.
By Amara Osei
3 min read

What's News
- LangChain launched LangSmith Tool Gateway in public beta on November 19, 2026.
- A large US retailer cited by LangChain runs more than 50 MCP servers exposing over 1,000 tools to agents.
- Acting-as-user and acting-as-agent modes separate employee credentials from service credentials.
- The beta covers LangSmith Plus and Enterprise plans across GCP US, GCP EU, GCP APAC and AWS US; Azure US is excluded.
LangChain put LangSmith Tool Gateway into public beta on November 19, 2026, giving enterprises a managed layer that decides which tools AI agents can call, whose credentials they use and how each action gets recorded.
Tool Gateway routes agent requests to MCP tool servers through a shared control layer on the LangSmith Agent Platform. It centralizes authentication, access rules and call records that teams would otherwise rebuild separately into each agent. LangChain is aiming at a production reality where agents serve users with different permissions, tokens need refreshing, and security teams must record access for audits.
How big is the problem it solves?
LangChain pointed to a large US retailer running more than 50 MCP servers that expose over 1,000 tools to store-operations and inventory agents. A store assistant must act under each associate's credentials across inventory, schedules and email, while an inventory agent reconciles stock at 3 a.m. under service credentials.
With hard-coded credentials or full server access, there are no deterministic checks preventing use of the wrong identity or an unapproved tool. Security teams also lack an end-to-end view for auditing, LangChain said.
What does the gateway actually control?
Administrators register a remote MCP server once or enable a LangSmith-hosted built-in server. They then assemble selected tools from different servers into a kit exposed through a single MCP URL. The kit determines which tools the agent can see. For large catalogs, a search mode lets the agent query the kit and load only task-relevant tools instead of the entire set.
Caller identity follows two models:
- Acting-as-user applies the end user's credentials, with Okta or Entra ID identifying the person. An unconnected user receives a sign-in link, and the call continues after authentication.
- Acting-as-agent supports unattended work through the agent's own credentials, or a profile holding shared credentials connected once by a workspace administrator.
The split lets a scheduled background process run under service permissions while an employee-facing assistant keeps the employee's permissions.
Where does human review kick in?
The Tool Gateway documentation fixes the order for every request: identify the user, agent or profile; apply deny policies; attach a stored credential or request a connection; check approval policies; forward the call. A trace is recorded when tracing is enabled.
Deny rules can target users, service keys or combinations, at organization or workspace level, including rules for everyone with listed exceptions.
Approval policies add a human checkpoint in the MCP client before a sensitive call executes. LangChain's example is a warehouse-writing tool: an administrator can require a person to confirm each attempted write. That places review at the moment of action rather than relying only on a later audit.
What can security teams see?
With tracing enabled, a call's record can include the tool, server, user, outcome and duration within the surrounding agent conversation. The data sent to or returned by the tool is not stored in that record. Enterprise plans add an organization-wide audit log showing which tool ran and under whose identity — a consolidated view of agent activity for security and compliance teams.
LangSmith-managed sign-in integrations cover Google, Microsoft, Salesforce, GitHub, Linear and LinkedIn. API-key or public built-ins include Parallel search, Exa, Tavily, Apollo, Ashby, Pylon and web-page reading. Tool kits work with MCP clients including LangChain, Deep Agents, Managed Deep Agents, Claude Code and Cursor.
Who gets access?
The beta runs on LangSmith Plus and Enterprise plans. On LangSmith Cloud, support covers GCP US, GCP EU, GCP APAC and AWS US; Azure US is excluded. Enterprise customers additionally get the organization-wide audit log.
The release also hands LangChain adjacent governance points for agent execution. LLM Gateway already controls model calls with features such as spending limits and fallbacks, while Tool Gateway governs interactions with external tools. For enterprises, that pairing pulls model usage and tool actions into one LangSmith control plane — with the beta's plan tiers and regional support setting the initial limits on deployment.
Source: unite.ai
More from Amara Osei
Show full bio
Senior reporter covering companies, deals and strategy in entrepreneurship, startups and venture funding at Business Bearings.
2,370 articles