Back to vendors
L

Langflow

Also known as: Langflow LFX

Visit site
Entry priceFree and open source under MIT, to self host. Enterprise support and managed hosting are priced by IBM and the hosting partners.Full pricing detail

Low-code open-source builder for RAG and agentic pipelines with visual flow editing, Python extensibility, and OSS MIT licensing.

Langflow is an open-source, low-code platform for visually building AI agents and retrieval-augmented generation applications. Developers connect components on a drag-and-drop canvas, each one representing a model, vector store, document loader, embedding, tool, or piece of custom logic, to assemble complex AI pipelines without writing all the wiring by hand. It is built in Python and uses LangChain under the hood.

Every component exposes its underlying Python source, so a developer can inspect and customize any node without leaving the interface. The canvas supports the major language model providers, including OpenAI, Anthropic, Google, Cohere and Hugging Face, as well as local models through Ollama, and vector stores come as provider bundles. An interactive Playground lets builders test and refine a flow step by step before shipping it.

A finished flow can be deployed as an API or exported as JSON for use in a Python application. Through built in MCP servers any workflow can become a tool that MCP compatible clients can call, and Langflow can also act as an MCP client itself. It supports multi-agent orchestration with conversation management and retrieval, and each component's output and logs can be inspected after a run. A desktop app for Windows and macOS bundles everything needed to start without managing Python environments. Langflow 1.8 added global model provider configuration, a new workflow API, and the MCP server and client support.

Langflow began as a project at Logspace, was acquired by DataStax in 2024, and came under IBM when IBM acquired DataStax. The hosted DataStax Langflow cloud version was retired in 2026 in favor of the open-source project, which remains freely available, self-hostable, and actively maintained, now positioned alongside IBM's watsonx data and AI portfolio.

Vendor details

Canonical URL

https://www.langflow.org/

Category

Agent builder

Subcategory

Open-source visual builder for LangChain-style flows, agents and RAG pipelines

Funding status

Acquired. Began as a project at Logspace, acquired by DataStax in 2024, and passed to IBM when IBM acquired DataStax. IBM now publishes a Langflow product page and positions it alongside watsonx, including deploying Langflow-built workflows into watsonx Orchestrate. The project remains MIT-licensed open source and independently versioned. IBM Langflow Cloud is reported by third parties to be in private preview, which no first-party page confirms.

Company status

acquired

Use cases & customers

Primary use cases

Building RAG pipelines and chat assistants visually from typed components, with the Python source of every node editable in placeComposing agents that call other flows and components as tools, with per-tool human approval gatesPublishing a finished flow as a REST endpoint, an embeddable chat widget, or an MCP server that external assistants call as a toolRunning the whole stack inside the customer's own boundary, including local models via Ollama, under an MIT licensePrototyping on a desktop application without managing Python environments, then exporting the flow as JSON for production

Target customers

developersSMB

Deployment options

SaaSself-hosted

Integrations

Components are organized as Core components grouped by purpose and Bundles grouped by service provider. Covered classes include language models across OpenAI, Anthropic, Google, Cohere and Hugging Face plus local models via Ollama, embedding models, vector stores as provider bundles, document loaders, text splitters, data sources and search providers, plus third-party bundles such as Apify actors. Generic escape hatches are an API Request component for any HTTP endpoint and fully custom Python components. Tool Mode converts any component into an agent-callable tool. Model Context Protocol works in both directions. Langflow connects outward as an MCP client to external MCP servers, and exposes its own flows as MCP servers that external clients such as IDE assistants can call as tools. Outbound surfaces are a REST API, JSON flow import and export, an embeddable langflow-chat web component, and LFX for embedding Langflow as a dependency.

In practice

You want to prototype a RAG chatbot fast but still need to tweak how a component actually works. Langflow lets you drag nodes together on a canvas and open the Python source of any one to customize it.

Your team is already in Python and LangChain and dislikes black-box builders. Langflow exposes every component's source code, so you build visually without giving up control over what each node does.

You've built a flow and want an AI assistant to call it as a tool. Langflow can deploy that workflow as an MCP server, turning it into a callable tool for any MCP-compatible client.

Agentic Index coverage score

12.0 / 14 capabilities · 86%

Integrations & Tool Calling Full

Components are organized as Core components grouped by purpose and Bundles grouped by service provider, covering language models, embeddings, vector stores, document loaders, data sources and tools.

An API Request component calls arbitrary endpoints, custom Python components extend the catalog arbitrarily, and Tool Mode on any component exposes it to an Agent component as a callable tool. As an MCP client Langflow connects to external MCP servers and uses their tools.

Provider bundles include Apify actors and search providers, and vector database and model provider bundles span the major commercial and local options.

Sourcedocs.langflow.org components overview, bundles, MCP client and API request pagesread 2026-08-31

Workflow Orchestration Full

Flows are node graphs of typed components with loop and branching behavior, parsing and type conversion between incompatible ports, and dynamic ports that open as templates reference new variables. Components can be grouped into a single reusable component and saved to the component menu, and freezing a component pins it and all upstream components so their output is reused rather than recomputed. Agents can use components and flows as tools, so one agent can invoke another flow, and individual components can be run in isolation to test dependencies.

Sourcedocs.langflow.org agents, components overview and flows pagesread 2026-08-31

Knowledge Grounding & RAG Full

Retrieval is assembled from component classes, document loaders, text splitters such as the Recursive Character Text Splitter with configurable chunk size and overlap, embedding model components, and vector store components supplied as provider bundles, so the customer chooses and owns the store. Retrieval components feed context into prompt templates through typed ports, with a Type Convert component bridging incompatible types. A file management system stores the names, paths, sizes and storage providers of uploaded files, and chat memory can itself be backed by an external vector store.

Sourcedocs.langflow.org components overview, vector store bundles, memory management and file management pagesread 2026-08-31

Human Oversight & Guardrails Full

Human-in-the-Loop pauses a flow, creates a checkpoint and waits for a human decision, after which the run resumes from the checkpoint along the selected branch without re-executing completed steps. It works two ways. The Human Input component pauses the flow where it is placed and creates one output branch per configured user action such as Approve or Reject, and per-tool approval on the Agent component, set by enabling Requires approval on an individual tool, pauses the run only when the agent attempts to call that tool. Human-in-the-Loop arrived in the Langflow 1.11 release.

Sourcedocs.langflow.org human-in-the-loop, human-input component and agent tools pagesread 2026-08-31

Security, Identity & Governance Partial

Langflow supports API key and JWT authentication, external authentication through an upstream OIDC proxy or corporate SSO gateway with JWKS validation and automatic user provisioning, and a pluggable authorization layer with viewer, developer and admin roles, shares and an authorization audit log. The open-source edition registers a pass-through authorization service and ships no enforcement plugin, so per-resource RBAC is not enforced without a third-party plugin. No security attestation applies to the self-hosted open-source project, and access enforcement is left to the operator.

Sourcedocs.langflow.org authorization and external-authentication pagesread 2026-09-29

Observability & Auditability Full

Each component has an Inspect control showing its output and logs for the most recent run, and a Last Run indicator confirms execution, so a builder can see what every node in the graph produced. Monitor endpoints expose stored messages at GET /v1/monitor/messages, filterable by session ID. The Playground shows message logs per session with full history, and log level and file settings are configurable through environment variables. Langflow does not describe tracing integrations with LangSmith or Langfuse.

Sourcedocs.langflow.org components overview, API reference and playground pagesread 2026-08-31

Memory & State Persistence Full

Memory bases store an agent's long-term chat history in a vector store for semantic retrieval of past conversations across sessions, filling themselves from each attached flow's runs through auto-capture and a configurable ingestion threshold, one memory base per flow and several per flow allowed, on Chroma, Chroma Cloud, OpenSearch or pgvector. Chat conversations are also stored by session ID, with the Agent component's chat memory on by default and a Message History component with store and retrieve modes, and message logs are editable and deletable.

Sourcedocs.langflow.org memory-bases and session-id pagesread 2026-09-29

Deployment & Data Residency Full

Langflow is an MIT-licensed open-source Python application that the customer self-hosts, deployable in containers that serve flows over a network, installable as a Python package, or run as Langflow Desktop, a downloadable application for macOS and Windows that bundles dependencies. LFX builds Langflow as a dependency within another application. Storage is a database the operator controls, LANGFLOW_USE_NOOP_DATABASE disables persistence entirely for testing, and chat memory can be directed to external databases the customer already runs.

Sourcedocs.langflow.org deployment overview, LFX overview, memory management pages and langflow.org desktopread 2026-08-31

Prebuilt Agents, Templates & Packs Full

Several pre-built templates come ready to use or customize, and many tasks begin by creating a flow from a named template such as Basic Prompting or the Memory Chatbot starter example. New workspaces open with a Starter Project folder by default. Grouped components can be saved into the Core components menu as reusable custom components, and flows are importable and exportable as JSON so complete assets can be shared and adopted.

Sourcedocs.langflow.org get started, memory, web-search and import and export pagesread 2026-08-31

Triggers & Channel Coverage Full

A Webhook component gives each flow its own POST /v1/webhook endpoint, authenticated by API key by default, so an event in an external application starts a run, and flows can also run through the API. Channels include the REST API, an embeddable langflow-chat web component, the Playground and flows published as MCP tools. Langflow has no built-in scheduler or cron trigger.

Sourcedocs.langflow.org llms.txt index and sitemap (webhook and component-webhook pages)read 2026-09-29

Model Flexibility & Routing Full

A Language Model component selects provider and model per node, so users add an OpenAI key or pick a different provider and model. Model providers come as Bundles grouped by provider, covering the major commercial options including OpenAI, Anthropic, Google, Cohere and Hugging Face, alongside local models through Ollama. The component can emit either a model response or a Language Model output that powers a downstream component's reasoning, so one flow can drive several models. Custom Python components allow any additional provider to be added.

Sourcedocs.langflow.org components overview, model components and web-search pagesread 2026-08-31

APIs, SDKs & MCP Extensibility Full

Langflow exposes flows as MCP servers that external MCP clients can use, and it also acts as an MCP client connecting to external MCP servers. A REST API runs flows, lists components at /api/v1/all, triggers runs and returns monitor data, authenticated by API key. Flows import and export as JSON, custom components are written in Python and can be shared, an embeddable langflow-chat web component brings flows into a site with host URL and API key props, and LFX builds Langflow as a dependency.

Sourcedocs.langflow.org MCP, API reference, import and export, custom components and publish pagesread 2026-08-31

Testing, Debugging & Optimization Partial

For testing and debugging, Langflow offers the Playground for interactive flow testing with custom session IDs, running a single component in isolation via build_vertex without executing upstream dependencies, Inspect for per-component output and logs, Freeze to keep a component and its upstream from re-running while preserving last output, message logs that can be edited or deleted to change later chatbot behavior, and automatic backup flows before component updates that may contain breaking changes. Langflow offers no evaluation, scoring, regression or benchmarking capability.

Sourcedocs.langflow.org playground, components overview and memory pagesread 2026-08-31

Browser & Computer Use Not documented

No browser, desktop or computer-use component ships with Langflow. The closest capability is the Web Search component, which combines web, news and RSS modes and uses web scraping subject to rate limits, alongside an API Request component for arbitrary HTTP calls. Browser capability reaches Langflow only through third parties, an Apify bundle that calls a hosted RAG Web Browser actor, and community-authored custom components wrapping external browser-automation libraries.

Sourcedocs.langflow.org web-search and components referenceread 2026-08-31

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Product updates

·Observability / auditabilityVerified

Langflow 1.13 adds opt in audit logging, an append only history of flow and project operations. Each event records who attempted a change, on which resource, when, how it completed and what was written. Once switched on, events can be queried through separate API endpoints for flows and for projects.

Bears on: Observability / auditability

View source
·Security / enterpriseVerified

Langflow 1.12.4 is a security hardening release with breaking changes for self hosted installs. Database connectors now block requests to private hosts unless they are allowlisted, database TLS files must sit in an admin controlled folder, MCP stdio configs can no longer pass Node.js runtime options, and chat memory and credentials are scoped more tightly to flow owners.

Bears on: Security / enterprise

View source
·Memory / stateVerified

Langflow released version 1.11 Desktop, introducing multi-vector retrieval capabilities. The update includes the lfx-nextplaid extension bundle, which provides native support for ColBERT late interaction and ColPali visual document retrieval.

Bears on: Memory / state

View source
View all 8 Langflow updates →Tracked since May 2026 · Verified from public vendor sources

Pricing

Free and open source under MIT, to self host. Enterprise support and managed hosting are priced by IBM and the hosting partners.

There is no Langflow subscription. Cost is the infrastructure, model and vector database usage the software runs on, paid to those providers, plus any IBM support contract.

Free tier

Included quota

The self hosted open source software is free under MIT with no limits. It includes a drag and drop canvas, 150+ prebuilt components, multi agent and RAG flows, native LangGraph support, MCP server export, and deployment as a REST API, an MCP endpoint or Python code. Enterprise support through IBM Elite Support for Langflow adds help backed by an SLA. Managed hosting through IBM watsonx, Render and Elestio is priced by those providers.

What is public

Langflow is an open source, low code visual builder for AI agents and RAG, built around LangChain. DataStax acquired it in 2024, it is now part of IBM's watsonx portfolio, and it is committed to stay "forever open, free, agnostic". It is MIT licensed and free to self host with no limits. The DataStax hosted Langflow Cloud was deprecated in March 2026 and shut down in April 2026, so there is no Langflow managed cloud tier. Managed hosting runs through partners (IBM watsonx, Render, Elestio), and enterprise support is through IBM Elite Support, priced by quote. The real cost is the infrastructure and the LLM and vector usage paid to those providers.

Billing mechanics

The software is free under MIT, and the cost is what it runs on. Self hosting covers the customer's own infrastructure, LLM API usage (OpenAI, Anthropic and others, passed through per token), vector databases (Pinecone, Weaviate) and observability. There has been no Langflow subscription floor since the DataStax hosted cloud shut down. Managed options bill through partners, such as Elestio's hourly credit model, and enterprise support is an IBM contract priced by quote.

Cost watchouts

The free software still carries real costs. Infrastructure, LLM tokens, the vector database and observability make up the bill. The shutdown of the DataStax hosted cloud in April 2026 forced a workflow migration to self hosting or partners, with data deletion. Flows assume LangChain abstractions, which locks them to LangChain. Releases ship monthly and occasionally bring breaking changes.

Variable cost rationale

Langflow itself meters nothing, because the software is free, so 100% of spend is variable. That spend is infrastructure, LLM tokens, the vector database and observability, paid to those providers, and it grows with deployment size and query volume. No Langflow cap applies, so the only limits are the ones the customer sets with its own providers.

Additional watchouts

The hosted cloud shut down in April 2026, so running Langflow means self hosting or a partner. LangChain lock-in makes a move to raw SDKs or another framework a rewrite. Large canvases (40+ nodes) get unwieldy, community component quality varies, fast monthly releases can bring breaking changes, and the builder is not pure no code.

Overage / add-ons

Langflow places no caps on the open source software. Cost scales with the customer's own infrastructure, LLM token usage and vector database usage, all paid to third parties. Managed hosting partners bill by resource and by the hour.

Sales call required

Mixed (some tiers require a call)

Free / trial

Free and open source under MIT, self hosted

Lowest paid plan

Not published

Commercial notes

Langflow launched in 2020 (Rodrigo Nader, Gabriel Almeida) and has 138,000+ GitHub stars. DataStax acquired it in April 2024, and it is now part of the IBM watsonx portfolio. It centers on LangChain, for RAG and agent prototypes, with MCP export.

Key ambiguities

Pricing is driven entirely by usage and infrastructure, with no Langflow subscription floor. Managed hosting is in flux after the DataStax cloud shutdown and the move to IBM. Enterprise support through IBM Elite Support is quoted, and managed hosting prices vary by provider.

Cancellation / refund

The self hosted software is free under MIT with no contract. IBM Elite Support for Langflow is a quoted contract, and partner hosting such as Elestio is pay as you go by resource with no lock in.

Support SLA / resale

The open source software has community support. IBM Elite Support for Langflow, formerly DataStax Luna, provides enterprise support backed by an SLA, and Langflow integrates with IBM watsonx Orchestrate as middleware.

Missing data

No Langflow subscription has existed since the hosted cloud shut down. Enterprise support and managed hosting prices depend on the provider and are quoted. The open source project is free under MIT and self hostable. Managed hosting runs through partners (IBM watsonx, Render, Elestio), and enterprise support is through IBM Elite Support, priced by quote.

Agentic Index verified 2026-09-29

Alternatives to Langflow

The closest documented capability profiles to Langflow among agent builders tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

  • AgentX12.0 / 14Fuller documented coverage on Testing, Debugging & Optimization
  • Airia12.0 / 14Fuller documented coverage on Testing, Debugging & Optimization
  • AutoGPT13.0 / 14Adds documented Browser & Computer Use
  • Dify12.0 / 14Fuller documented coverage on Security, Identity & GovernanceLangflow vs Dify →
  • LuMay12.0 / 14Fuller documented coverage on Security, Identity & Governance
  • Odin AI13.0 / 14Adds documented Browser & Computer Use

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.