[Langflow 1.9 just released!](/content/blog/langflow-1-9/index.html) Langflow OSS 1.9 is released! This version introduces new capabilities for building, operating, and integrating AI workflows, with updates focused on in-product AI assistance, flow deployment tooling, and MCP-based interoperability. **Core improvements** - Langflow Assistant - Flow DevOps Toolkit - MCP support for IDEs and coding agents - Token usage display **Also included:** - `Data` renamed to `JSON` and `DataFrame` renamed to `Table` - Flow versioning - Environment variables to block custom component execution - Gemini 3 tool calling support - Policies component (beta) - …and more Let's dig into some of these features to see how they make Langflow easier to integrate with external AI systems. ## Langflow Assistant Langflow Assistant is an AI assistant embedded directly into Langflow, like an additional Playground. It allows you to generate custom components from natural language, troubleshoot flows, and get real-time guidance directly from the docs. Langflow Assistant turns Langflow into a more interactive AI-assisted development environment, where help is available where you're building, instead of in a separate terminal or documentation tab. For more information, see [Build components with Langflow Assistant](https://docs.langflow.org/langflow-assistant). ## Flow DevOps Toolkit Langflow 1.9 introduces the Flow DevOps Toolkit SDK for versioning, testing, and deploying flows directly from your terminal. Initialize a Langflow project with `lfx init YOUR_PROJECT_NAME`, and the Flow DevOps Toolkit SDK creates a complete scaffold for deploying flows from local to production servers. Control deployments with the `environments.yaml` file, version flows, and push and pull changes between your Langflow server and local environments. For more information, see [Flow DevOps SDK toolkit](https://docs.langflow.org/flow-devops-sdk). ## MCP support for IDEs and coding agents Langflow now allows IDEs and coding agents to use Langflow programmatically through standardized MCP protocols. Tools such as IBM Bob and Claude Code can build and execute flows in Langflow directly, making Langflow a stronger part of agentic development workflows. For more information, see [Langflow MCP client for coding agents](https://docs.langflow.org/langflow-mcp-client). ## Token usage display LLM components now expose input and output token counts directly in the flow interface after execution for immediate visibility into token consumption. ## `Data` renamed to `JSON` and `DataFrame` renamed to `Table` The `Data` type and port are renamed to `JSON`. The `DataFrame` type and port are renamed to `Table`. Flows using `Data` and `DataFrame` are backwards compatible. For more information, see [Langflow data types](https://docs.langflow.org/data-types). ## Flow versioning Flows can now be versioned from the sidebar. Save flows in incrementing versions, or revert changes back to a previous version. For more information, see [Build flows](https://docs.langflow.org/concepts-flows). ## Environment variables to block custom component execution (beta) The new `LANGFLOW_ALLOW_CUSTOM_COMPONENTS` environment variable allows Langflow administrators to disable execution of custom Python components. This improves security in controlled deployments. Use it in conjunction with `LANGFLOW_COMPONENTS_PATH` to create an allow-list while still blocking other custom code execution. For more information, see [Block custom component execution](https://docs.langflow.org/deployment-block-custom-components). ## Gemini 3 tool calling support Langflow 1.9 adds tool-calling compatibility for Gemini 3 models. Support for Gemini tool calling required additional handling for Gemini's unique `thought_signature` field, and it is now supported in Langflow workflows. ## Policies component (beta) Langflow 1.9 adds a Policies component powered by [ToolGuard](https://github.com/AgentToolkit/toolguard). The Policies component converts natural language business policies into executable guard code, enabling runtime validation over tool execution without requiring custom rule scripting. For more information, see [Policies component](https://docs.langflow.org/policies). ## There's more to come For more on all the changes that have been released in Langflow OSS 1.9, you can check out the [release notes](https://docs.langflow.org/release-notes#190) or the [full release on GitHub](https://github.com/langflow-ai/langflow/releases/tag/v1.9.0). We're committed to continuously improving Langflow, and your input is crucial to this process. If you want to get more involved you can: - Join our [Discord community](https://discord.com/invite/EqksyE2EX9) to network with other developers - Explore our [GitHub repository](https://github.com/langflow-ai/langflow) to contribute to the project or 🌟 the project - Subscribe to our [YouTube channel](https://www.youtube.com/@langflow) for step-by-step guides on using Langflow - Stay updated on the latest AI trends and insights by subscribing to the [AI++ newsletter](/content/newsletter/index.html) We're excited to see the projects you'll create with the Langflow 1.9 open-source package. Check out the [1.9.0 release on GitHub](https://github.com/langflow-ai/langflow/releases/tag/v1.9.0) and get building! To download Langflow Desktop 1.9, see [Download Langflow Desktop](/content/desktop/index.html).