The latest AI news and updates of LangChain — AI company building frameworks and observability tools for developing and orchestrating LLM agents. Covering LangChain's latest product updates, launches, and analysis from the past 90 days.
LangChain40m agoLangChain added cache hit rate tracking to LangSmith, providing visibility into how often agent LLM calls land on cached results. The update includes an overall hit rate stat, per-thread breakdowns, and time-series dashboard templates. These metrics surface caching efficiency, as cached LLM calls reduce latency and token costs compared to standard requests.
LangChain2h agoLangChain open-sourced Content Studio, an internal design agent that generates on-brand videos, graphics, and slide decks from Slack prompts. The agent uses a sandbox-per-thread harness, 35 specialized skills, and asynchronous subagents to manage complex design tasks. Teams can fork the repository to add brand tokens and deploy the agent to their own Slack workspaces using Managed Deep Agents.
LangChainOct 16LangChain detailed how Managed Deep Agents secures credentials within sandboxes. The system uses a proxy to inject user tokens outside the sandbox environment, restricting access to only pre-approved services. This design ensures that even if an agent executes malicious third-party packages, the underlying credentials remain isolated and invisible to the agent process.
LangChainOct 15LangChain integrated NVIDIA's open-weight Nemotron 3.5 Super VL model into its Deep Agents framework via ChatNVIDIA. The 120B-parameter model achieved an 87% pass rate on LangChain's behavioral suite. In internal Harbor-lite benchmarks, it completed tasks 55% faster than DeepSeek V4.1 Flash with similar success rates.
LangChainOct 13LangChain added support for ScaleDown's constrained decision models, which are optimized for quick, cost-effective tasks like summarization, extraction, and classification. These models integrate with LangChain and Deep Agents, allowing every call to be traced through the LangSmith LLM Gateway.
LangChainOct 13LangChain launched AI Search in LangSmith, enabling teams to query agent traces using natural language to identify specific behaviors like user frustration or task failures. The feature uses a decision model to analyze agent trajectories, running alongside existing filters. It is available in public beta on LangSmith Cloud across GCP US, EU, and APAC regions at no additional cost.
LangChainOct 13LangChain partnered with ScaleDown to integrate constrained decision models for classification, extraction, and summarization into agent workflows. These task-specific models skip token-by-token generation to improve decision speed and cost efficiency. The integration is available through the new langchain-scaledown OSS package and the LangSmith LLM Gateway, which adds tracing, spend limits, and rate limits for production deployments.
LangChainOct 9LangChain released Managed Deep Agents 0.8, simplifying agent authentication, memory, and communication channels. The update introduces user-owned credentials, user-level memory, and Slack file transfers. Additionally, Managed Deep Agents now includes built-in web search powered by Parallel, allowing builders to configure search servers directly within project files without managing separate vendor accounts or API keys.
LangChainOct 9LangChain published a case study detailing Snyk Assist, a conversational agent built on LangChain and LangGraph. Since April, the agent has handled over 60,000 queries across 500 customer accounts, resolving 85%+ of sessions without a support ticket. It also auto-detected and escalated 250 cases to the appropriate teams, with observability managed through LangSmith.
LangChainOct 8LangChain introduced a pattern for building agents that handle real-world purchases using Stripe's Link wallet and the Machine Payments Protocol. The workflow runs in Slack, where agents search for products and build carts, while human-in-the-loop approval ensures every payment is authorized through the Link wallet. This design keeps sensitive payment credentials isolated from the model.
LangChainOct 8LangChain revamped skills in its Deep Agents framework to improve context efficiency. The update adds tool binding, which loads tools only when a skill is read, and allows apps to pin skills at runtime. Agents can now reload skills mid-thread to pick up library changes without starting a new conversation.
LangChainOct 8LangChain released Managed Deep Agents v0.9, introducing a Schedules SDK for agent-created reminders and recurring tasks. The update adds per-run configuration, allowing a single deployment to select models, skills, and tools for each specific execution. Additionally, agents now automatically post Slack reactions upon starting a run to acknowledge incoming messages.
LangChainOct 6LangChain integrated web search into Managed Deep Agents, powered by p0. The update removes the need for separate vendor accounts, API key management, or manual tool wiring. Builders can now configure the search server directly within their project files to augment agent responses with external web data.
LangChainSep 28LangChain launched LangSmith Engine v2, introducing Red Teaming to proactively detect hallucinations and system-prompt violations before production. The update adds detection for inefficient agent trajectories and performance trends in latency and cost. Engine v2 automatically validates proposed fixes by reproducing failures and testing solutions against real inputs before providing a one-click PR for deployment.
LangChainSep 25LangChain launched LangSmith Custom Apps, providing a way to build and publish custom interfaces directly on agent data within the LangSmith workspace. The feature generates UIs from prompts or templates, removing the need for external hosting, authentication, and permissions. Plus plans include one app per organization, while Enterprise plans offer unlimited access to the custom interface builder.
LangChainSep 24LangChain launched LangSmith Fine-Tuning and the smithtune CLI, now in Public Beta. The workflow converts existing agent trajectories into specialized fine-tuned models, managing dataset preparation, training on open models via Baseten and Fireworks, and evaluation within LangSmith. This process creates specialized models that reduce inference cost and latency compared to general-purpose frontier models.
LangChainSep 24LangChain launched Trajectories in LangSmith, a chronological, conversational view of agent sessions. The feature aggregates human, AI, and tool messages across main agents and subagents, presenting them in order of appearance to simplify debugging and review. Trajectories are available now on all LangSmith plans in the US.
LangChainSep 24LangChain launched Managed Deep Agents 0.8, adding user-owned credentials and user-level memory for caller-specific context. The release introduces HTTP channels for internal tool integration, native Slack file transfers, and built-in web search powered by Parallel. These features provide managed infrastructure for production agents, handling authentication, memory, and communication channels without custom boilerplate code.
LangChainSep 23LangChain now supports cron-based scheduling for Managed Deep Agents. By adding a schedule file with a cron expression and a prompt to the project directory, the agent executes autonomously on a set interval. This feature is currently in public beta on LangSmith Cloud in the US region, supporting ephemeral threads and automated Slack delivery for final responses.
LangChainSep 22LangChain demonstrated Patch, an agent built on Managed Deep Agents that automates code changes. When tagged in a Slack thread with a feature description, the agent generates the necessary code and opens a GitHub pull request with the diff. Managed Deep Agents provides the agent harness and production infrastructure for these file-based agent projects.
LangChainSep 21LangChain added TypeSafe AI's Jev model as a judge in LangSmith, enabling fast, typed evaluations of agent traces. In internal tests, Jev matched human reviewers on every call, averaging 0.44 seconds and $0.34 per run, compared to $28.17 for Claude Sonnet 4.6. The model evaluates multiple criteria in parallel, allowing teams to score every production trace.
LangChainSep 20LangChain benchmarked TypeSafe AI’s Jev model as an agent evaluator, finding it matched a human oracle on 100% of 500 binary decisions. Jev achieved 92–913x lower quality-score variance than frontier LLM judges while averaging 0.44s and $0.00035 per call. This decision-first model returns typed answers directly, offering a stable, low-cost alternative to traditional LLM-as-a-judge methods.
LangChainSep 17LangChain released Deep Life Sci, an open source agentic assistant for clinical and lab scientists. Built on the Deep Agents harness, the tool accesses over 600,000 clinical trial records and 41 million scientific articles. It includes a LangSmith sandbox for data analysis and delegates tasks to sub-agents to process large document sets for research and compliance workflows.
LangChainSep 17LangChain rebuilt the filtering experience in LangSmith to improve agent trace analysis. The update introduces precise query building to narrow down runs, tools to locate specific executions, and logic to explain why a result matched a filter. These features support drilling into agent behavior and debugging complex multi-step workflows.
LangChainSep 16LangChain now enables custom middleware for Managed Deep Agents, providing hooks to inject logic around model calls, tool calls, and the agent lifecycle. This update allows for deterministic policy enforcement, such as PII redaction, dynamic tool selection, and context management. Developers can now implement bespoke behaviors to customize agent harnesses for specific production use cases.
LangChainSep 16LangChain updated its Deep Agents framework to include built-in context management, addressing context window overload in long-running tasks. The framework now ships with integrated filesystems, subagents, and skills to manage information flow. These components allow agents to offload data and delegate tasks, maintaining a clean context window for complex, multi-step operations.
LangChainSep 15Lyft reduced customer support AI agent development from six months to 1-2 weeks by building a self-serve platform with LangGraph and LangSmith. The architecture supports agent definition through prompts and JSON configurations, removing technical intermediaries from daily iteration. Automated LLM-as-a-judge evaluation pipelines and production monitoring maintain agent reliability and safety.
LangChainSep 15LangChain updated Managed Deep Agent deployments to serve as MCP servers via a built-in /mcp endpoint. This allows coding assistants and other agents to invoke deployed agents as tools for research and analysis. The feature is currently available in public beta on LangSmith Cloud for the US region, supporting stateless requests via Streamable HTTP transport.
LangChainSep 14LangChain now enables Managed Deep Agents to connect to Slack, allowing agents to receive and respond to messages directly within Slack conversations. The integration supports app mentions, direct messages, and thread replies. This feature is currently available in public beta on LangSmith Cloud for the US region.
LangChainSep 14LangChain updated LangSmith LLM Gateway to support model-specific API key governance. This feature locks an API key to a set of permitted models, automatically blocking any unauthorized model calls. The update is currently available in Public Beta for LangSmith Plus and Enterprise plans, enforcing consistent model access and usage limits across agent infrastructure.
LangChainSep 10LangChain integrated the Harbor framework into Managed Deep Agents to provide reproducible evaluation environments. Each agent evaluation now runs in a fresh container, with results automatically tracked in LangSmith. This integration allows developers to verify that skill additions, model swaps, or tool description tweaks maintain agent performance without breaking existing functionality.
LangChainSep 9LangChain integrated Browserbase’s Stagehand SDK into its Deep Agents framework, providing browser tools without custom definitions. The integration ships as an MCP server offering run, snapshot, and screenshot capabilities. It requires a persistent ClientSession and uses stdio transport to isolate dependency environments, supporting agent interaction with web pages through Browserbase or local Chrome.
LangChainSep 9LangChain launched Connections in Managed Deep Agents 0.7.0+, moving credential management from local .env files to the LangSmith workspace. This update introduces agent-owned secrets for shared tool access and user-owned OAuth, which resolves credentials per caller at runtime. Agents now automatically handle OAuth flows and carry the identity of the specific user requesting the action.
LangChainSep 4LangChain analyzed LangSmith observability data from the 14 days ending August 30, 2026, to track model usage. gpt-4o-mini led in organizational adoption at 13%, while gpt-4.1-mini drove 7% of total call volume. DeepSeek V4 Flash ranked second in adoption at 9%, marking the only open-weight model to appear in the top-10 lists for reach or volume.
LangChainSep 3LangChain updated its Python MCP adapters to support the new stateless Model Context Protocol specification. The client now defaults to a stateless architecture, where each tool invocation creates a fresh session, while maintaining support for stateful sessions via explicit configuration. These updates also include improved error handling for tool execution failures.
LangChainAug 30LangChain published a new Max Agency podcast episode featuring Unify CTO Connor Heggie. The discussion details how Unify reduced AI agent compute costs by 90–95% two weeks before launch. Engineering fixes included optimizing prompt caching, replacing full virtual machines with a suspending Python REPL, and implementing subagents as function calls to prevent runaway budget depletion.
LangChainAug 26LangChain published a case study detailing Rippling's production AI evaluation pipeline, which uses four distinct layers to ensure system health. The architecture includes offline mock-based evals, post-merge integration tests, deploy-blocking scenarios, and continuous production monitoring. Rippling uses LangSmith to manage these traces and analyze conversation data at scale across its multi-agent system.
LangChainAug 26LangChain launched WikiBench, a benchmark for evaluating codebase documentation agents on the Harbor framework. Testing shows that combining a wiki with raw source code improves accuracy and cost-efficiency for coding agents compared to using source code alone. The benchmark uses a reader agent to score wiki quality based on grounded answers to repository-specific questions.
LangChainAug 25LangChain updated LangSmith Engine to deliver over twice the performance on internal benchmarks for identifying agent issues. The platform now supports self-hosted deployments and includes a Reduced Analysis mode for cost-sensitive users. Additionally, the tool adds Slack and Linear integrations for issue tracking and automatically closes stale issues that no longer appear in production traces.
LangChainAug 25Toyota North America scaled its enterprise AI portfolio from one production agent every six months to over 50 agents in four days. The team built this platform using LangChain’s Deep Agents and LangGraph, while utilizing LangSmith as an observability board to monitor agent performance, tool failures, and pipeline regressions across the manufacturing line.
LangChainAug 22LangChain launched LangSmith Tuned Evaluators to automatically score agent behavior in production, starting with Perceived Error. These managed evaluators replace manual prompt and infrastructure maintenance, delivering frontier-level accuracy at an 82% lower cost. The Perceived Error evaluator analyzes threads with at least two message pairs and is available now for Plus and Cloud Enterprise plans in the US.
LangChainAug 22LangChain launched LangSmith Preview Builds in public beta for LangSmith Deployment. This feature creates temporary, production-like environments from pull request branches, allowing teams to test agent changes, inspect traces, and share feedback before merging. Previews are short-lived and automatically tear down after the pull request merges, streamlining the agent review process without extra CI/CD setup.
LangChainAug 12LangChain launched LangSmith Bring Your Own Cloud on AWS, providing a managed deployment within a customer's own VPC. This model keeps sensitive agent traces, datasets, and runtime data inside the customer's AWS boundary while LangChain handles infrastructure provisioning, scaling, and upgrades. BYOC is available for Enterprise customers across 15 AWS regions in the US, EU, and APAC.
LangChainAug 12LangChain rebuilt LangSmith Dashboards to enhance investigation and reporting workflows. The update introduces free-form layout arrangement, KPI widgets alongside trend charts, and cross-unit metric comparisons. Dashboards now support trace breakdowns by model or user, custom notes, and configurable charts for tracking specific application metrics.
LangChainAug 11LangChain benchmarked NVIDIA’s Switchyard router on 145 multi-step agent tasks, finding 93% of turns could be handled by a 30B model instead of a frontier model. This routing strategy reduced total costs by approximately 70% while retaining 90% of frontier-level accuracy. LangChain released a Deep Agents integration to enable this routing in production agent workflows.
LangChainAug 8LangChain launched Managed Deep Agents in public beta, providing a hosted runtime for production-ready agents. The service handles durable execution, persistence, sandboxes, and memory, allowing teams to deploy agents via CLI without managing underlying infrastructure. It includes first-class support for Harbor evals and Slack channels, enabling focus on agent behavior and business logic.
LangChainAug 8LangChain’s LangSmith Signal report finds that one in three active teams ran an open-weight model in July 2026, up from one in five a year ago. Daily usage of open-weight models grew 9x year-over-year, outpacing the 4.8x growth seen in closed models. This analysis draws from production observability data across billions of agent runs on the LangSmith platform.
LangChainAug 6LangChain added Moonshot AI's Kimi K3 and K2.6 models to the LangSmith LLM Gateway. Powered by Fireworks AI, the integration allows model access using a LangSmith API key and Gateway Credits, eliminating the need for separate provider accounts. Usage is governed by existing workspace spend policies and includes full LangSmith tracing for production observability.
LangChainAug 5LangChain updated OpenWiki with a built-in web visualizer, accessible via the openwiki visualize command. The tool serves a local interface with an interactive node graph and a live Markdown reader for exploring repository documentation and file relationships. The visualizer automatically updates as wiki files change during the session.
LangChainAug 5LangChain released an autonomous SRE agent for Kubernetes that performs proactive health checks and on-demand diagnostics. The agent uses a scheduler to bypass full model calls for routine checks, reducing costs by 95–99%. It features a human-in-the-loop safety model for all write actions and uses LangSmith Engine to automatically detect and fix issues in production traces.
LangChainAug 4LangChain published a framework for evaluating voice agents across three distinct dimensions: Execution, Outcome, and Experience. The approach measures instruction adherence, goal achievement, and conversational smoothness separately to identify specific agent weaknesses. Evaluations use LangSmith to combine code evaluators, LLM judges, and business-system checks for monitoring performance and tracking improvements over time.
LangChainAug 3Stripe built Kai, its company-wide Knowledge AI Platform, in one week using LangChain’s Deep Agents harness. The agent connects to internal data, Slack, and Google Suite, reaching over 5,000 users in four weeks. It currently supports 1,000 skills and 500 internal MCP tools, with 83% of Stripe employees using it weekly for data analysis and document drafting.
LangChainJul 31LangChain released ReviewBench, a benchmark designed to evaluate code review performance based on real pull-request feedback. The system curates actual review issues into reproducible tasks that run on the Harbor framework. This provides a standardized way to measure how effectively AI agents identify and address the specific types of issues human reviewers catch in production code.
LangChainJul 30LangChain released the LangSmith LLM Gateway in public beta, providing a centralized governance layer for production AI agents. The gateway enforces spend and rate limits, configures model fallback policies, and redacts sensitive data like PII and secrets before requests reach model providers. It supports multiple providers and custom endpoints to help manage agent costs and reliability.
Frequently asked questions