# LangChain — Open source agent frameworks plus LangSmith, a hosted platform to trace, evaluate, deploy and govern agents. > Source: The Agents Index — https://theagentsindex.com/langchain (structured, researched, re-verified) > Facts last verified: 2026-09-18 LangChain publishes open source agent frameworks (LangChain, LangGraph and the dcode terminal coding agent) and sells LangSmith, a hosted platform for tracing, evaluating, deploying and governing those agents. Fleet adds no-code agent building on the same platform. | Fact | Value | | --- | --- | | Website | https://www.langchain.com | | Price | from $39/seat per month | | Free tier | Yes | Developer plan: $0 per seat, maximum of 1 seat, up to 5k base traces per month then pay-as-you-go, 5 LCU of Fleet usage and 5 LCU plus 1 LSU of sandbox usage per month, limited to 10 sandboxes, community support, no Deployment and no Engine. ## What it does **Writes code** · Yes — The vendor ships dcode, a free open-source terminal coding agent that writes and runs code from a prompt; it is a separate package from the paid LangSmith platform. > dcode is an open-source terminal coding agent built on the Deep Agents SDK. Bring your own model, customize the agent harness, and control how code execution is approved, traced, and run. > — https://www.langchain.com/dcode · Hero section, 'The coding agent you own' **Debugs and fixes** · Partly — LangSmith Engine diagnoses failures from production traces and opens a GitHub pull request with the fix, but a human still merges it, and Engine is absent from the free Developer plan. > Engine reads your source code to locate the code path behind a failing trace, ground its proposed fixes in the actual implementation, and open pull requests directly from issues. > — https://docs.langchain.com/langsmith/engine · Set up Engine, step 'Connect a code repository (optional)' **Writes tests** · Partly — Engine generates ground-truth dataset examples and assertions so a fix can be evaluated offline; those are agent evals, not unit tests, and the vendor's CI guide leaves unit tests to you. > Tracks the issue over time, automatically adding new traces that match the same pattern, and generates ground truth dataset examples so you can verify a fix. > — https://docs.langchain.com/langsmith/engine · How Engine works, step 4 **Whole-codebase work** · Partly — Engine reads a whole connected repository to ground fixes and dcode carries context across sessions, but no page claims a repository-wide migration or multi-file refactor as a supported job. > Engine can propose code changes to any connected repository, including agents built with Deep Agents, LangChain, and LangGraph. > — https://docs.langchain.com/langsmith/engine · Take action on an issue, 'Open or view a pull request' **Ships it** · Yes — LangSmith Deployment runs agents in production with one-click deploy from Studio and a control-plane API; the free Developer plan shows Deployment as N/A, so shipping starts at Plus. > Go from testing graphs locally in Studio to deploying them on Langsmith Cloud in 1 Click, directly from Studio. > — https://docs.langchain.com/langsmith/studio · Deploy from Studio section **Done-for-you service** · Partly — You run the software yourself on every self-serve plan; only Enterprise buyers get LangChain engineers, architectural guidance and team training alongside the product. > Access to deployed engineers > — https://www.langchain.com/pricing · Plan comparison table, Support section, 'Access to deployed engineers' row (Enterprise column only) _Not established: Reviews code_ ## How much it does unattended **Runs autonomously** · Yes — Engine scans traces and drafts fixes on a schedule with nobody watching, and deployed agents run as background or cron-scheduled jobs rather than only in a chat session. > Engine is scheduled to run once every 6 hours. > — https://www.langchain.com/pricing · FAQ, LangSmith Engine Questions, 'How often does Engine run?' **Multi-agent** · Yes — LangGraph is built for coordinating several agents in one graph, including hierarchical supervisors, and dcode delegates work to task-specific subagents in parallel. > Design diverse control flows — single, multi-agent, hierarchical — all using one framework. > — https://www.langchain.com/langgraph · How does LangGraph help?, 'Build expressive, customizable agent workflows' **Agent permissions** · Yes — Human-in-the-loop interrupts gate agent actions in the framework, dcode gates code execution behind approval modes, and Fleet agents can run as themselves or as the calling user. > Add human-in-the-loop checks to steer and approve agent actions. > — https://www.langchain.com/langgraph · How does LangGraph help?, human-in-the-loop section ## Which models it runs on **Claude** · Yes — A first-party langchain-anthropic package covers Claude with tool calling, structured output and streaming, and the LLM Gateway routes Anthropic model IDs with one LangSmith key. > You can find information about Anthropic's latest models, their costs, context windows, and supported input types in the Claude docs. > — https://docs.langchain.com/oss/python/integrations/chat/anthropic · ChatAnthropic integration, opening paragraph **GPT** · Yes — A first-party langchain-openai package covers GPT models and Azure OpenAI deployments, and the gateway accepts OpenAI Chat Completions and Responses requests for them. > You can find information about OpenAI's latest models, their costs, context windows, and supported input types in the OpenAI Platform docs. > — https://docs.langchain.com/oss/python/integrations/chat/openai · ChatOpenAI integration, opening paragraph **Open models** · Yes — Open-weight models run through local Ollama, vLLM, Hugging Face and Together integrations, and the gateway can front a self-hosted OpenAI-compatible endpoint you operate yourself. > Ollama allows you to run open-source Large Language Models (LLMs), such as gpt-oss, locally. > — https://docs.langchain.com/oss/python/integrations/chat/ollama · ChatOllama integration, opening paragraph **Model choice** · Yes — You choose the provider everywhere except Fleet's Fast, Pro and Max tiers, where LangChain selects and maintains the underlying model for you; custom configurations can override even those. > All features support custom workspace configurations, so you can use any provider or model—even for features that show a curated list by default. > — https://docs.langchain.com/langsmith/model-configurations · Feature Access section, note under the table ## Where you use it **In your editor** · Partly — LangSmith Studio is the vendor's own agent IDE for visualising and debugging graphs, but no plugin for VS Code, JetBrains or Cursor is documented; dcode is a terminal, not an editor. > Studio is a specialized agent IDE that enables visualization, interaction, and debugging of agentic systems that implement the Agent Server API protocol. > — https://docs.langchain.com/langsmith/studio · Opening paragraph **On the command line** · Yes — Several terminal surfaces exist: the LangGraph CLI for local dev servers, a Sandbox CLI on every plan, a LangSmith CLI for issues, and the dcode terminal coding agent. > You can list issues programmatically using the LangSmith CLI. > — https://docs.langchain.com/langsmith/engine · List issues via the CLI section **In your pipeline** · Yes — The vendor documents a GitHub Actions pipeline that runs tests and offline evals on a pull request, then promotes preview deployments to production through the control-plane API. > This guide demonstrates how to implement a comprehensive CI/CD pipeline for AI agent applications deployed in LangSmith Deployment. > — https://docs.langchain.com/langsmith/cicd-pipeline-example · Opening paragraph **In a browser** · Yes — Everything paid runs in a hosted console at smith.langchain.com (traces, evals, Engine issues, Fleet agents and billing), and Fleet agents also reach users in Slack, Teams and Gmail. > In the LangSmith console, select Engine in the UI sidebar. > — https://docs.langchain.com/langsmith/engine · Set up Engine for a tracing project, step 1 ## Whose machine it runs on **Self-hosted** · Partly — The open source frameworks run anywhere, but self-hosting or hybrid-hosting the LangSmith platform on Kubernetes is an Enterprise-plan option; Developer and Plus are cloud-only. > Self-hosted and hybrid deployment options > — https://www.langchain.com/pricing · Plans, Enterprise plan card bullet list **Open source** · Partly — The frameworks and the dcode coding agent are MIT-licensed and free, while the LangSmith platform that carries the price, seats and governance is proprietary and hosted. > Yes - LangChain is an MIT-licensed open-source library and is free to use. > — https://www.langchain.com/langchain · FAQs for LangChain, 'Is LangChain open source?' ## What it costs to run **How it meters** · Yes — Billing mixes a per-seat licence with pay-as-you-go consumption: traces beyond the included allowance, plus compute units for Engine, Fleet, deployments and sandboxes. > LangChain Compute Unit (LCU): A LCU is a normalized unit of work done on the LangSmith Platform and is used for pricing purposes. > — https://www.langchain.com/pricing · FAQ, General Questions, 'What is an LCU? What is an LSU?' **Free tier** · Yes — The Developer plan is a standing $0 tier with one seat and a monthly trace allowance, not a countdown trial, and the open source packages cost nothing at all. > Up to 5k base traces / mo, then pay-as-you-go > — https://www.langchain.com/pricing · Plans, Developer plan card bullet list **API access** · Yes — The LangSmith REST API is open to every plan and Fleet agents can be triggered by API, but the Agent Server's assistants and state endpoints arrive with the Plus plan. > 30+ API endpoints including state and memory > — https://www.langchain.com/pricing · Plan comparison table, LangSmith Deployment section (Plus and Enterprise columns checked) **MCP server** · Yes — Agents can be exposed as MCP servers on every plan including the free one, and can themselves consume tools from remote MCP servers; LangSmith also publishes its own MCP server. > Expose agent as MCP server > — https://www.langchain.com/pricing · Plan comparison table, LangSmith Deployment section (Developer, Plus and Enterprise columns checked) **Bring your own key** · Partly — You can bring your own provider keys through the gateway on every plan, or buy inference as Gateway Credits, but Engine refuses your keys and bills its model use as compute units. > Engine uses LangChain-managed inference exclusively. Bring Your Own Key (BYOK) is not supported; you cannot supply your own provider API keys for Engine. > — https://docs.langchain.com/langsmith/engine · Configure Engine section, BYOK note ## Buying it for a team **A company can buy it** · Yes — Plus is bought self-serve by an organisation with unlimited seats and shared workspaces; Enterprise adds annual invoicing, custom terms and an infosec review for procurement. > Add unlimited seats $39 per seat/month > — https://www.langchain.com/pricing · Plan comparison table, Pricing section, Users row (Plus column) **Seat model** · Yes — Seats are counted per person including pending invitations, the free plan is capped at one seat, and Plus has no seat floor or ceiling at $39 each per month. > A seat is a distinct user inside your organization. We consider the total number of users (including invited users) to determine the number of seats to bill. > — https://www.langchain.com/pricing · FAQ, General Questions, 'What is a seat?' **Pooled budget** · Yes — Fleet's compute allowance is one balance for the whole organisation that resets monthly, and Engine spend limits are set org-wide or per project rather than per seat. > Developer & Plus plans include a free usage allowance, shared across your organization and resets monthly > — https://www.langchain.com/pricing · FAQ, LangSmith Fleet Questions, 'How is Fleet metered?' **Admin controls** · Yes — Admins cap model spend and request rates through the gateway on every plan and set monthly compute ceilings on Engine; fine-grained role and tag policies need Enterprise. > Central governance: Apply spend limits, rate limits, and data policies. > — https://docs.langchain.com/langsmith/llm-gateway · What the gateway provides section **Audit log** · Partly — A tamper-resistant log of administrative actions with a UI, an API and 400-day retention exists, but the vendor states plainly it is Enterprise-only: Developer and Plus get nothing. > No. Audit logs are an Enterprise feature. > — https://docs.langchain.com/langsmith/audit-logs · FAQ, 'Are audit logs available on the Plus or Developer plan?' **Single sign-on** · Partly — Google and GitHub social login come with the cheap plans, but SAML single sign-on and SCIM provisioning are Enterprise-only; self-hosted installs can wire up almost any OIDC provider. > Enterprise customers can configure SAML SSO and SCIM. > — https://docs.langchain.com/langsmith/authentication-methods · Cloud section, SAML SSO subsection ## What happens to your code **Opt out of training** · Yes — No opt-out is needed because the vendor states it never trains models on customer traces, prompts or outputs; Engine's analysis still runs through LangChain-managed inference. > No. LangSmith does not use your data to train models. Your traces, prompts, and outputs remain private to your organization. > — https://www.langchain.com/pricing · FAQ, General Questions, 'Will you train on the data that I send to LangSmith?' **Data residency** · Yes — Cloud customers pick a US or EU region, hybrid keeps the data plane in your own VPC, and a self-hosted install keeps everything inside your infrastructure. > Host LangSmith in LangSmith's managed cloud with US or EU data residency. > — https://docs.langchain.com/langsmith/enterprise · Hosting options, Cloud card **Getting out** · Partly — Traces can be queried through the SDK on any plan, but scheduled bulk export to your own S3 bucket in Parquet is now Enterprise-only for new customers, with older accounts cut off in 2027. > For customers who signed up after August 3, 2026, bulk export is only available on the LangSmith Enterprise plan. > — https://docs.langchain.com/langsmith/data-export · Plan restrictions callout at top of page **Certifications** · Yes — The platform claims SOC 2 Type II, GDPR and HIPAA with annual third-party audits and penetration testing, plus AES-256 at rest and TLS 1.2 or higher in transit. > We maintain SOC 2 Type II, GDPR, and HIPAA compliance and undergo annual third-party audits and penetration testing. > — https://docs.langchain.com/langsmith/shared-responsibility-model · LangChain responsibilities (the platform), first bullet ## Platform record (as of 2026-08-30) Reported by the code-hosting platform, not by the vendor. | Fact | Value | | --- | --- | | Licence | MIT | | Stars | 145277 | | Forks | 24250 | | Last push | 2026-08-30 | ## Provenance 1 of 31 sources behind the facts above are independent of the vendor. ## Coverage elsewhere Pages about this product that the vendor does not control. The label states how independent each one is. | Source | Kind | Read | | --- | --- | --- | | fortune.com | Independent source | 2025-10-20 | | techcrunch.com | Independent source | 2025-10-21 | ## What changed - 2026-09-18 — Capabilities · Commercial terms · Free tier - 2026-09-02 — Capabilities · Commercial terms · Free tier · Price - 2026-08-31 — Capabilities · Commercial terms · Free tier · Price - 2026-08-19 — One-liner - 2026-08-17 — Tag:claude ## Correcting this record Anything marked "Not established" is a question we have not answered — it is not a No. If you represent LangChain, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth. Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.