Create a new Langflow workflow with optional configuration and organization.

Purpose: Initialize a new flow in your Langflow instance. Use this to create workflows from scratch or with predefined structure. The flow can be organized into folders and configured with initial data including nodes, edges, and component settings.

Parameters:

• name (required, string, max: 255 chars): The display name of the flow (must be unique within folder)

• description (optional, string): Brief explanation of flow's purpose and functionality

• data (optional, object): Initial flow configuration containing nodes, edges, and component definitions

• folder_id (optional, UUID string): ID of parent folder for organization (use list_folders to get valid IDs)

Returns: FlowRead object containing:

• id (UUID): Newly created flow's unique identifier (save this for future operations)

• name (string): Flow display name

• description (string): Flow description

• folder_id (UUID): Parent folder if specified

• user_id (UUID): Owner user identifier

• created_at, updated_at (ISO timestamps): Creation and modification timestamps

• data (object): Complete flow structure with nodes and edges

Usage Examples:

1. Create minimal flow: { name: "My Workflow" }

2. Flow with description: { name: "Data Pipeline", description: "Processes user data" }

3. Flow in folder: { name: "API Flow", folder_id: "folder-uuid-here" }

4. Flow with structure: { name: "Chat Bot", data: { nodes: [...], edges: [...] } }

Best Practices:

• Use descriptive names that indicate flow purpose (e.g., "Customer Onboarding" not "Flow1")

• Add descriptions to document flow functionality for team collaboration

• Organize related flows using folder_id for better project management

• Start with minimal configuration, then use update_flow to add complexity

• Validate data structure before passing complex flow configurations

Common Errors:

• "Flow name already exists": Choose a unique name or specify different folder_id

• "Invalid folder ID": Ensure folder_id is valid UUID from list_folders

• "Name is required": Must provide name parameter

• "Data validation failed": Check that data object follows Langflow schema structure

• "Unauthorized": Verify LANGFLOW_API_KEY is valid

Related Tools:

• list_flows: View all created flows

• update_flow: Modify flow after creation

• get_flow: Retrieve full flow details by ID

• upload_flow: Import flow from JSON file instead of creating from scratch

• create_folder: Create folders before organizing flows

List Langflow flows with flexible filtering and pagination.

Purpose: Retrieve flows from your Langflow instance with optional pagination and organization filters. Essential for discovering available workflows, managing large flow collections, and browsing flows within specific folders.

Parameters:

• page (optional, number, default: 1, min: 1): Page number for pagination

• size (optional, number, default: 50, max: 100): Number of flows per page

• folder_id (optional, UUID string): Filter flows within a specific folder

• components_only (optional, boolean, default: false): Return only component flows (reusable flow components)

• get_all (optional, boolean, default: true): Retrieve all flows ignoring pagination (use with caution for large datasets)

Returns: Array of FlowRead objects containing:

• id (UUID): Unique flow identifier

• name (string): Flow display name

• description (string): Flow purpose description

• folder_id (UUID): Parent folder if organized

• user_id (UUID): Owner user identifier

• created_at, updated_at (ISO timestamps): Flow lifecycle dates

• data (object): Flow configuration with nodes, edges, and component definitions

Usage Examples:

1. List recent flows (default): {}

2. Paginate through flows: { page: 2, size: 20 }

3. Flows in specific folder: { folder_id: "uuid-here" }

4. Get all component flows: { components_only: true, get_all: true }

Best Practices:

• Use pagination (page/size) for better performance instead of get_all

• Filter by folder_id when working within organized projects

• Keep page size <= 50 for optimal response times

• Cache folder_id values to reduce API calls

• Use get_all sparingly, only when you need complete dataset

Common Errors:

• "Page size cannot exceed 100": Reduce size parameter to 100 or less

• "Invalid folder ID format": Ensure folder_id is a valid UUID (use list_folders to get valid IDs)

• Empty results: Check if folder_id exists or try removing filters

• Connection timeout: Reduce page size or enable pagination instead of get_all

Related Tools:

• create_flow: Create new flows after listing existing ones

• get_flow: Retrieve complete flow details by ID from this list

• list_folders: Find valid folder_id values for filtering

• update_flow: Modify flows found in this list

• delete_flow: Remove flows identified in this list

Retrieve complete details and configuration of a specific flow.

Purpose: Fetch comprehensive information about a flow including its full node graph, component configurations, connections, and metadata. Essential for inspecting flow structure, debugging workflows, and preparing flows for modification or execution.

Parameters:

• flow_id (required, UUID string): The unique identifier of the flow to retrieve (obtained from list_flows or create_flow)

Returns: FlowRead object containing:

• id (UUID): Flow's unique identifier

• name (string): Flow display name

• description (string): Flow purpose description

• folder_id (UUID): Parent folder identifier if organized

• user_id (UUID): Owner user identifier

• created_at, updated_at (ISO timestamps): Flow lifecycle timestamps

• data (object): Complete flow structure including:

  • nodes (array): All component nodes with their configurations

  • edges (array): Connections between nodes

  • viewport (object): Canvas positioning data

  • description (string): Flow-level documentation

Usage Examples:

1. Get flow details: { flow_id: "550e8400-e29b-41d4-a716-446655440000" }

2. Inspect before modification: { flow_id: "flow-uuid-from-list" }

3. Debug flow structure: { flow_id: "failing-flow-uuid" }

Best Practices:

• Use list_flows first to discover available flow_id values

• Cache flow data locally to reduce API calls when repeatedly accessing same flow

• Check updated_at timestamp to detect if flow changed since last retrieval

• Inspect data.nodes to understand flow's component structure

• Review data.edges to trace data flow between components

• Validate flow_id format (must be valid UUID) before making request

Common Errors:

• "Flow not found": flow_id doesn't exist or was deleted (verify with list_flows)

• "Invalid UUID format": flow_id must be valid UUID string (e.g., "550e8400-e29b-41d4-a716-446655440000")

• "Unauthorized": API key lacks permission to access this flow

• "Flow ID is required": Must provide flow_id parameter

• Connection timeout: Check Langflow instance availability and network connectivity

Related Tools:

• list_flows: Find available flow_id values

• update_flow: Modify flow configuration after retrieval

• run_flow: Execute flow after inspecting its structure

• build_flow: Compile and validate flow structure

• delete_flow: Remove flow after inspection confirms it's no longer needed

Modify existing flow properties, configuration, or organization.

Purpose: Update flow metadata (name, description), modify flow structure (nodes, edges, components), or reorganize flows into different folders. Essential for maintaining flows, fixing configurations, renaming workflows, and managing project organization over time.

Parameters:

• flow_id (required, UUID string): Unique identifier of flow to update (from list_flows or get_flow)

• name (optional, string, max: 255 chars): New display name for the flow

• description (optional, string): Updated purpose and functionality description

• data (optional, object): New flow structure with nodes, edges, and component configurations

• folder_id (optional, UUID string): Move flow to different folder (use list_folders for valid IDs, null for root level)

Returns: FlowRead object containing:

• id (UUID): Flow's unique identifier (unchanged)

• name (string): Updated flow name

• description (string): Updated description

• folder_id (UUID): New folder location

• user_id (UUID): Owner identifier

• created_at (ISO timestamp): Original creation time (unchanged)

• updated_at (ISO timestamp): Timestamp of this update

• data (object): Complete updated flow structure

Usage Examples:

1. Rename flow: { flow_id: "flow-uuid", name: "Improved Customer Bot" }

2. Update description: { flow_id: "flow-uuid", description: "Processes customer inquiries with sentiment analysis" }

3. Move to folder: { flow_id: "flow-uuid", folder_id: "folder-uuid" }

4. Move to root: { flow_id: "flow-uuid", folder_id: null }

5. Update structure: { flow_id: "flow-uuid", data: { nodes: [...], edges: [...] } }

6. Rename and move: { flow_id: "flow-uuid", name: "New Name", folder_id: "folder-uuid" }

Best Practices:

• Get current flow data with get_flow before updating to avoid overwriting changes

• Only specify fields you want to change (partial updates supported)

• Rebuild flow with build_flow after data modifications to validate changes

• Use descriptive names and descriptions for team collaboration

• Update description when flow functionality changes significantly

• Move flows to appropriate folders to maintain organization

• Verify folder_id exists before moving (use list_folders)

• Update flow name if purpose or functionality changes

Common Errors:

• "Flow not found": flow_id doesn't exist or was deleted (verify with list_flows)

• "Flow name already exists": Choose unique name within target folder

• "Invalid folder ID": folder_id is not valid UUID (use list_folders to get valid IDs)

• "Folder not found": Specified folder_id doesn't exist

• "Invalid flow data structure": data object doesn't match Langflow schema

• "Concurrent modification": Flow was updated by another user (re-fetch with get_flow)

• "flow_id is required": Must provide flow_id parameter

• "Unauthorized": API key lacks permission to modify this flow

Related Tools:

• get_flow: Retrieve current flow state before updating

• build_flow: Validate flow after data modifications

• list_flows: Find flow_id for flows to update

• list_folders: Get valid folder_id values for reorganization

• create_flow: Create new flow instead of modifying existing one

Permanently remove a single flow from Langflow instance.

Purpose: Delete flows that are no longer needed, removing obsolete workflows, or cleaning up test/development flows. This operation is destructive and irreversible - deleted flows cannot be recovered unless previously backed up via download_flows.

Parameters:

• flow_id (required, UUID string): Unique identifier of flow to delete (from list_flows or get_flow)

Returns: Success object containing:

• success (boolean): true if deletion succeeded

• message (string): Confirmation message

Usage Examples:

1. Delete single flow: { flow_id: "550e8400-e29b-41d4-a716-446655440000" }

2. Remove test flow: { flow_id: "test-flow-uuid" }

Best Practices:

• Download flow backup with download_flows before deletion if you might need it later

• Verify flow_id is correct using get_flow before deleting to avoid mistakes

• Check if flow is used in production or referenced by other systems before deletion

• Document why flow was deleted for audit purposes

• For deleting multiple flows, use delete_flows (batch operation) instead

• Review flow contents one final time before permanent removal

• Consider archiving flows to dedicated folder instead of deleting

• Ensure no running jobs or active sessions using this flow

Common Errors:

• "Flow not found": flow_id doesn't exist or already deleted (verify with list_flows)

• "Invalid UUID format": flow_id must be valid UUID string

• "Cannot delete running flow": Flow has active execution (wait for completion or cancel)

• "flow_id is required": Must provide flow_id parameter

• "Unauthorized": API key lacks permission to delete this flow

• "Flow is referenced": Flow is used as component in other flows (remove references first)

Related Tools:

• download_flows: Backup flow before deletion for potential recovery

• list_flows: Find flow_id for flows to delete

• get_flow: Inspect flow before deletion to confirm it's the right one

• delete_flows: Delete multiple flows at once (more efficient for bulk operations)

Delete multiple flows at once. Pass array of flow IDs.

List all available Langflow components. Returns components with their types, names, and descriptions.

Execute a Langflow flow with inputs and retrieve execution results.

Purpose: Run a workflow and receive its outputs. Essential for testing flows, integrating workflows into applications, and automating tasks. Supports both synchronous execution and streaming for real-time responses from LLM-based flows.

Parameters:

• flow_id_or_name (required, string): Flow identifier (UUID) or flow name (case-sensitive)

• input_request (required, object): Execution configuration containing:

  • input_value (optional, string): Primary input data for the flow (e.g., user message, data to process)

  • output_component (optional, string): Specific output component to return

  • output_type (optional, string): Expected output format (e.g., "chat", "text", "json")

  • input_type (optional, string): Type of input being provided (e.g., "chat", "text", "json")

  • session_id (optional, string): Session ID for conversation continuity

  • tweaks (optional, object): Component-specific parameter overrides (key: component_id, value: parameter object)

• context (optional, object): Request context passed alongside the input request

• stream (optional, boolean, default: false): Streaming is currently rejected by this MCP client because streamed responses are not JSON serializable

Returns: RunResponse object containing:

• outputs (array): Flow execution results from output components

• session_id (string): Unique session identifier for conversation continuity

• Additional fields vary by flow configuration (message, data, artifacts)

Usage Examples:

1. Simple execution: { flow_id_or_name: "my-chatbot", input_request: { input_value: "Hello" } }

2. With output type: { flow_id_or_name: "data-processor", input_request: { input_value: "data", output_type: "json" } }

3. With context: { flow_id_or_name: "llm-chat", input_request: { input_value: "Explain AI" }, context: { tenant: "acme" } }

4. With tweaks: { flow_id_or_name: "flow-uuid", input_request: { input_value: "test", tweaks: { "component-id": { temperature: 0.7 } } } }

Best Practices:

• Use flow ID (UUID) for production; names may change or conflict

• Build flow first with build_flow to ensure it's valid before execution

• Leave stream false; this MCP client currently returns JSON tool responses only

• Cache session_id for multi-turn conversations to maintain context

• Set appropriate timeout values for long-running flows

• Test flows with various inputs before production deployment

• Use tweaks sparingly; prefer configuring flows in Langflow UI

Common Errors:

• "Flow not found": Verify flow_id_or_name exists using list_flows

• "Flow has not been built": Run build_flow before executing

• "Invalid input configuration": Ensure input_request matches flow's expected inputs

• "Component configuration error": Check tweaks object matches component IDs and parameter names

• "Execution timeout": Flow took too long; increase timeout or optimize flow

• "Missing required input": Flow expects input_value but none provided

Related Tools:

• build_flow: Compile flow before execution to validate structure

• get_flow: Inspect flow structure to understand required inputs

• get_build_status: Check build completion before running flow

• trigger_webhook: Alternative execution method via webhook endpoint

• list_flows: Find available flows to execute

Trigger a flow via webhook endpoint. Simplified API for webhook-based flow execution.

Upload a flow from file data. Provide an object with name, content (stringified JSON), and optional type fields.

Download flows as JSON export. Exports multiple flows for backup or transfer.

Get basic example flows. Returns a list of pre-built example flows for learning and templates.

List all folders with pagination for organizing flows hierarchically.

Purpose: Retrieve folders from your Langflow instance to discover organizational structure and obtain folder IDs for flow organization. Essential for understanding project hierarchy, finding folders for flow placement, and managing workspace organization.

Parameters:

• page (optional, number, default: 1, min: 1): Page number for pagination

• size (optional, number, default: 50, max: 100): Number of folders per page

Returns: Array of FolderRead objects containing:

• id (UUID): Unique folder identifier (use as folder_id in create_flow or update_flow)

• name (string): Folder display name

• description (string): Folder purpose description

• parent_id (UUID): Parent folder ID if nested, null for root-level folders

• created_at, updated_at (ISO timestamps): Folder lifecycle timestamps

Usage Examples:

1. List all folders (default): {}

2. Paginate through folders: { page: 2, size: 25 }

3. Get first 10 folders: { page: 1, size: 10 }

Best Practices:

• Call this before create_flow to find appropriate folder_id for organization

• Use pagination for large folder collections to improve performance

• Cache folder IDs and names to reduce API calls

• Check parent_id to understand folder hierarchy and nesting structure

• Keep page size <= 50 for optimal response times

• Map folder IDs to names for user-friendly displays in UIs

Common Errors:

• "Page size cannot exceed 100": Reduce size parameter to 100 or less

• "Invalid page number": page must be >= 1

• Empty results: No folders exist (create folders with create_folder)

• Connection timeout: Reduce page size or check Langflow instance availability

Related Tools:

• create_folder: Create new folders after reviewing existing organization

• get_folder: Retrieve detailed information about specific folder

• create_flow: Use folder_id from this list to organize new flows

• update_folder: Modify folder properties found in this list

• list_flows: Filter flows by folder_id obtained from this list

Create a new folder for organizing flows with optional nested hierarchy.

Purpose: Establish organizational structure in Langflow by creating folders to group related flows. Supports hierarchical organization through nested folders, enabling project-based or team-based flow management. Essential for maintaining organized workspace with many flows.

Parameters:

• name (required, string, max: 255 chars): Folder display name (should be descriptive and unique)

• description (optional, string): Purpose and contents description for documentation

• parent_id (optional, UUID string): Parent folder ID for creating nested subfolder hierarchy (use list_folders to find valid parent IDs)

Returns: FolderRead object containing:

• id (UUID): Newly created folder's unique identifier (use this as folder_id when creating/organizing flows)

• name (string): Folder display name

• description (string): Folder description

• parent_id (UUID): Parent folder ID if nested, null for root-level folder

• created_at, updated_at (ISO timestamps): Folder lifecycle timestamps

Usage Examples:

1. Create root folder: { name: "Production Flows" }

2. Folder with description: { name: "Customer Services", description: "Customer-facing automation workflows" }

3. Create nested subfolder: { name: "API Integrations", parent_id: "parent-folder-uuid" }

4. Project organization: { name: "Q1 2024 Project", description: "First quarter automation initiatives" }

Best Practices:

• Use descriptive, hierarchical names (e.g., "Marketing/Email Campaigns" structure via nesting)

• Add descriptions to document folder purpose and ownership

• Plan folder hierarchy before creation (root categories, then subcategories)

• Create folders before flows to enable immediate organization

• Use consistent naming conventions across team (e.g., "Team - Project - Category")

• Limit nesting depth to 2-3 levels for maintainability

• Cache created folder IDs for subsequent flow creation operations

Common Errors:

• "Folder name already exists": Choose unique name or create in different parent folder

• "Invalid parent folder ID": Ensure parent_id is valid UUID from list_folders

• "Parent folder not found": parent_id doesn't exist or was deleted

• "Name is required": Must provide name parameter

• "Maximum nesting depth exceeded": Too many nested levels (flatten hierarchy)

• "Circular reference detected": parent_id creates loop in folder structure

Related Tools:

• list_folders: Browse existing folders before creating new ones

• update_folder: Modify folder name, description, or move to different parent

• create_flow: Create flows organized within this folder using returned ID

• get_folder: Retrieve folder details after creation

• delete_folder: Remove folder when no longer needed

Get details of a specific folder by ID.

Update an existing folder. Can change name, description, or move to different parent.

Delete a folder by ID. Warning: This may affect flows in the folder.

List all projects with optional pagination. Projects group related flows and resources.

Create a new project. Projects help organize and group related flows.

Get details of a specific project by ID.

Update an existing project. Can change name or description.

Delete a project by ID. Warning: This may affect flows in the project.

Upload a project from file data. Provide an object with name, content (stringified JSON), and optional type fields.

Download a project as JSON export. Exports project for backup or transfer.

List all global variables. Variables store reusable values across flows.

Create a new global variable. Variables can be referenced in multiple flows.

Update an existing variable. Can change name, value, or type.

Delete a variable by ID. Warning: Flows using this variable may be affected.

Compile and validate a flow, returning job ID for asynchronous status monitoring.

Purpose: Build and validate a flow's component graph, ensuring all nodes are properly configured and connected. Essential prerequisite for flow execution - detects configuration errors, validates component parameters, and prepares the flow for running. Returns job_id for tracking build progress asynchronously.

Parameters:

• flow_id (required, UUID string): Flow identifier to build and validate

• inputs (optional, object): Input values to use during build validation (key: input_name, value: test_data)

• data (optional, object): Flow configuration override (use to build modified version without updating flow)

• files (optional, array of strings): File paths to include in build context

• stop_component_id (optional, string): Component ID where build should stop (useful for partial validation)

• start_component_id (optional, string): Component ID where build should start (skip earlier components)

• log_builds (optional, boolean, default: true): Enable detailed build logging for debugging

• flow_name (optional, string): Override flow name during build (doesn't affect saved flow)

• event_delivery (optional, string, default: "polling"): Delivery mode - "polling" (recommended), "streaming", or "direct"

Returns: BuildFlowResponse object containing:

• job_id (UUID): Unique build job identifier for status polling via get_build_status

Usage Examples:

1. Basic build: { flow_id: "550e8400-e29b-41d4-a716-446655440000" }

2. Build with logging: { flow_id: "flow-uuid", log_builds: true }

3. Build with test inputs: { flow_id: "flow-uuid", inputs: { "user_message": "test" } }

4. Partial build: { flow_id: "flow-uuid", start_component_id: "node-1", stop_component_id: "node-5" }

5. Streaming events: { flow_id: "flow-uuid", event_delivery: "streaming" }

Best Practices:

• Always build flow before first execution to catch configuration errors early

• Use polling event_delivery for reliable status tracking in production

• Enable log_builds during development for detailed error diagnostics

• Rebuild after any flow modifications (nodes, edges, parameters)

• Monitor build status with get_build_status using returned job_id

• Use partial builds (start/stop component) for faster validation of specific sections

• Validate inputs during build to ensure flow accepts expected data format

• Cache successful build status to avoid unnecessary rebuilds

Common Errors:

• "Flow not found": flow_id doesn't exist (verify with list_flows or get_flow)

• "Invalid flow configuration": Flow has disconnected nodes or missing required parameters

• "Component validation failed": Node configuration is invalid (check component settings in Langflow UI)

• "Missing required component parameters": Required fields not set on components

• "Build timeout": Complex flow took too long; simplify or optimize component chain

• "Invalid component ID": start_component_id or stop_component_id doesn't exist in flow

• "Circular dependency detected": Flow has loop in node connections

Related Tools:

• get_build_status: Poll build completion and retrieve validation results using job_id

• run_flow: Execute flow after successful build

• get_flow: Inspect flow structure before building

• cancel_build: Stop long-running build job

• update_flow: Fix configuration errors found during build

Get build status and events for a specific build job. Use polling mode to check async build progress.

Cancel a running build job. Stops the build process for the specified job.

Build a public flow without authentication using Langflow 1.11.x public-build semantics.

Get events for a public flow build job, including streaming event payloads.

Cancel a public flow build job.

List all knowledge bases for Retrieval-Augmented Generation (RAG) workflows.

Purpose: Discover available knowledge bases containing document collections for RAG applications. Essential for understanding what document repositories are available, finding knowledge bases for flow integration, and managing document-based AI applications.

Parameters: None

Returns: Array of KnowledgeBaseInfo objects containing:

• name (string): Knowledge base identifier/name (use this in get_knowledge_base or delete_knowledge_base)

• description (string): Purpose and contents description

• Additional metadata varies by knowledge base implementation (document count, vector store info, embedding model)

Usage Examples:

1. List all knowledge bases: {}

Best Practices:

• Call this before integrating knowledge bases into RAG flows to verify availability

• Use knowledge base names in Langflow's Vector Store or RAG components

• Document which flows use which knowledge bases for dependency tracking

• Monitor knowledge base list to identify unused repositories for cleanup

• Cache knowledge base names for flow configuration validation

• Verify knowledge base exists before flow execution to avoid runtime errors

Common Errors:

• Empty results: No knowledge bases configured (create knowledge bases in Langflow UI or via API)

• Connection timeout: Check Langflow instance availability and vector store connectivity

• "Knowledge base service unavailable": Vector store backend not configured or unreachable

Related Tools:

• get_knowledge_base: Retrieve detailed information about specific knowledge base

• delete_knowledge_base: Remove individual knowledge base when no longer needed

• bulk_delete_knowledge_bases: Clean up multiple knowledge bases at once

• create_flow: Build RAG flows using knowledge base names from this list

Get detailed information about a specific knowledge base by name.

Delete a specific knowledge base by name. Warning: This will permanently remove the knowledge base.

Delete multiple knowledge bases at once. Useful for batch cleanup operations.

Upload a file to a specific flow. Supports multipart/form-data for binary file uploads.

Download a specific file from a flow. Returns file content as base64.

List all files associated with a specific flow.

Delete a specific file from a flow.

Get an image file from a flow. Returns image content as base64.

List all available profile pictures.

Get a specific profile picture by folder and file name.

Validate Python code for custom components. Checks syntax and provides error/warning feedback.

Validate prompt template syntax. Checks for valid variable syntax and returns extracted variables.

Check if the Langflow component store is enabled and accessible.

Validate a store API key for accessing the Langflow component store.

Browse available components in the Langflow store. Supports pagination, filtering by tags, and search.

Get detailed information about a specific store component.

List all available component tags in the store.

Get components liked by the current user.

Advanced flow execution using Langflow 1.11.x input/output request arrays, tweaks, sessions, and streaming.

Execute a flow with session-based state management. Maintains conversation context across multiple calls using the same session ID.

Get current user registration status from Langflow. Returns whether the user is registered and their email if available.

Register a new user with Langflow using their email address.

Legacy process endpoint for backward compatibility. Processes a flow with inputs and optional tweaks.

Legacy predict endpoint for backward compatibility. Predicts flow output with inputs and optional tweaks.

Get build execution history for a specific flow. Retrieves vertex build map model containing build status and execution details. Essential for monitoring flow build history, debugging build failures, and tracking build performance over time.

Query chat/message history with flexible filtering options. Retrieve messages filtered by flow ID, session, sender, or sender name. Supports custom ordering for chronological or reverse-chronological message retrieval. Essential for debugging chat flows, analyzing conversation patterns, and monitoring user interactions.

Get detailed information about a specific message by ID. Retrieves complete message data including sender, content, timestamp, and metadata. Useful for inspecting individual message details and debugging message-related issues.

List all chat session IDs, optionally filtered by flow. Returns array of session identifiers that can be used to query session-specific messages. Essential for discovering active sessions, monitoring concurrent users, and analyzing session patterns.

Get all messages for a specific session ID. Retrieves complete conversation history for a session in chronological order. Essential for reviewing full conversation context, debugging session-specific issues, and analyzing user interactions within a session.

Migrate messages from old session ID to new session ID. Updates all messages from the old session to belong to the new session, effectively merging or renaming sessions. Useful for session continuity, combining split sessions, or correcting session identifiers.

List transaction logs for a flow with pagination support. Retrieves detailed transaction history including execution logs, API calls, and system events. Essential for debugging flow execution, analyzing performance bottlenecks, and monitoring API usage patterns.

Delete build history for a specific flow. Removes all stored build execution records to clean up old data and free storage space. Use cautiously as this permanently removes build history that may be useful for debugging.

Delete multiple messages by their IDs. Batch deletion of message records for cleanup, privacy compliance, or storage management. Permanently removes messages from the system.

⚠️ DEPRECATED: This endpoint is deprecated in Langflow API 1.6.4 and will be removed in a future version. Use "build_flow" tool instead with start_component_id/stop_component_id for partial builds.

Get vertex build order for a flow. Retrieves the ordered list of vertex IDs representing the build execution sequence. Useful for understanding flow component dependencies, optimizing build performance, and debugging build order issues. Supports partial builds with start/stop component IDs.

⚠️ DEPRECATED: This endpoint is deprecated in Langflow API 1.6.4 and will be removed in a future version. Use "get_build_status" tool with event_delivery="streaming" to monitor build progress.

Stream real-time build events for a specific vertex. Provides live updates during vertex build process using server-sent events (SSE). Essential for monitoring long-running component builds, debugging build failures in real-time, and providing live build status to users.

Get Langflow API version information. Retrieves current API version number and related system information. Essential for compatibility checks, debugging version-specific issues, and ensuring client-server version alignment.

List all users in the Langflow instance. Admin-only endpoint. Supports pagination with skip and limit parameters. Returns array of user objects with id, username, email, and other profile information.

Get information about the currently authenticated user. Returns user profile including id, username, email, is_active, is_superuser status, and profile image.

Update user profile information. Can modify username, password, or profile image. Returns updated user object.

Reset password for a specific user. Admin-only endpoint. Requires user ID and new password (minimum 8 characters). Returns updated user object.

List all API keys for the authenticated user. Returns array of API key objects with id, name, created_at, last_used_at, total_uses, and is_active status.

Create a new API key with a specified name. Returns the created API key object including the key value (only shown once).

Delete a specific API key by ID. This action is permanent and cannot be undone. The API key will immediately stop working.

Create a new custom component with Python code. Requires component code, name, and optional description and return_type. Returns created component object.

Authenticate with username and password. Returns access token, optional refresh token, and token type. Use the access token in subsequent API requests.

Automatically login with stored credentials or session. Returns access token if auto-login is configured and enabled.

Refresh authentication token using refresh token. Returns new access token and optionally a new refresh token.

Logout and invalidate current authentication session. Clears authentication tokens and terminates the session.

Get a public flow by ID without authentication. Public flows can be accessed by anyone without API key or login credentials.

Create multiple flows in a single batch operation. More efficient than creating flows individually. Accepts array of flow objects.

⚠️ DEPRECATED: This endpoint is deprecated in Langflow API 1.6.4 and will be removed in a future version. Use "get_build_status" tool with job_id instead of task_id.

Get status of an asynchronous task by task ID. Returns task status (pending, running, completed, failed), result if completed, or error if failed.

Download entire folder as compressed archive. Returns archive file containing all flows and resources in the folder.

Upload folder from compressed archive. Provide base64-encoded file content and filename. Creates folder structure and imports all flows.

List available starter project templates. Returns array of pre-built project templates that can be used to quickly bootstrap new Langflow projects.

Ingest a file into an existing knowledge base for RAG. Provide knowledge base name, base64-encoded file content, and filename.

List available ElevenLabs text-to-speech voices. Returns array of voice objects with voice_id and name for use in voice synthesis flows.

Check Langflow instance health status. Returns health status and system information. Use to verify the API is accessible and operational.

Retrieve system logs. Supports both regular and streaming modes. Use stream parameter to enable real-time log streaming.

Replace an entire flow (PUT) with new content. Overwrites name, description, data, and folder for the given flow ID.

Expand flows by resolving component references into full flow definitions. Pass the expansion request payload as the body object.

Get the event stream/history for a flow. Optionally filter to events since a given cursor/timestamp.

Create a new event for a flow (e.g. deployment or lifecycle event).

List saved versions of a flow with optional pagination and deployment provider filter.

Create a new saved version (snapshot) of a flow. Optionally provide a body with version metadata.

Get a specific saved version of a flow by version ID.

Delete a specific saved version of a flow.

Activate a specific saved version of a flow, making it the current version. Optionally save the current state as a draft first.

Detect global variables referenced by one or more flow versions.

Save the Langflow Store API key for the current user so store operations are authenticated.

Update a custom component template field, recomputing the frontend node based on the provided code and field changes.

Publish a new component to the Langflow Store.

Like (or toggle like on) a Langflow Store component.

Create a response via the OpenAI-compatible /responses endpoint using a model and input.

Get the current authenticated session information.

Create a new Langflow user account.

Get webhook events for a flow by ID or name. Optionally filter by user ID.

Get the detailed health check report from the Langflow server root /health_check endpoint.

List user-scoped files via the /api/v2/files endpoint.

Upload a user-scoped file via /api/v2/files. Provide base64 file content and filename; optionally append or mark ephemeral.

Get a user-scoped file by ID via /api/v2/files. Optionally return the file content.

Rename a user-scoped file via /api/v2/files.

Delete a user-scoped file by ID via /api/v2/files.

Delete ALL user-scoped files via /api/v2/files. This is irreversible.

Batch download multiple user-scoped files via /api/v2/files by their IDs.

Batch delete multiple user-scoped files via /api/v2/files by their IDs.

List knowledge bases with detailed metadata (embedding config, chunk counts, etc.).

Create a new knowledge base with an embedding provider/model and optional column configuration.

Preview how a file would be chunked for knowledge base ingestion. Provide base64 file content, filename, and optional chunking params.

List chunks stored in a knowledge base with optional pagination and search.

Ingest a file into an existing knowledge base. Provide base64 file content, filename, and optional ingestion params.

Cancel an in-progress knowledge base ingestion job.

Update a stored monitor message (text, sender, session, files, or properties).

Delete all monitor messages for a given session.

Delete multiple monitor sessions by their IDs.

Get shared (playground) messages for a source flow, optionally filtered by session and ordered.

Get the list of shared session IDs for a source flow.

Update a shared (playground) message for a source flow.

Migrate a shared session to a new session ID within a source flow.

Delete a shared session within a source flow.

List execution traces with optional filters (flow, session, status, query, time range) and pagination.

Delete all traces associated with a flow.

Get a single execution trace by ID.

Delete a single execution trace by ID.

List available models with optional filters (provider, name, type, capabilities, search).

List all known model providers.

List enabled model providers, optionally restricted to a given set of provider names.

List enabled models, optionally restricted to a given set of model names.

Enable or disable specific models by provider and model ID.

Get the default model for a given model type.

Set the default model (provider + model name) for a given model type.

Delete (clear) the default model for a given model type.

Get the mapping between model providers and their required global variables.

Validate a model provider configuration with the provided variables (e.g. API keys).

Get available language (LLM) model options.

Get available embedding model options.

Request agentic assistance for building or editing a flow component.

Request streaming agentic assistance for building or editing a flow component.

Check whether the agentic assistant is configured and available.

Execute an agentic assistant flow by name with the given assist request fields.

Get the result of a workflow run, optionally by job ID.

Run a Langflow 1.12.x v2 workflow with sync, stream, or background mode.

Stop a running workflow job by job ID.

List configured MCP servers. Optionally include action counts.

Get the configuration of a specific MCP server by name.

Create an MCP server configuration (stdio command or remote URL based).

Update an existing MCP server configuration by name.

Delete an MCP server configuration by name.

Get the MCP configuration for a project. Optionally filter to only MCP-enabled flows.

Update the MCP configuration (settings and auth settings) for a project.

Get the installation status of a project as an MCP server across clients.

Install a project as an MCP server into a target client, optionally specifying transport.

Get the MCP Composer URL for a project.

List authorization roles. Optionally filter by system flag or name with pagination.

Get a single authorization role by ID.

Create an authorization role with optional permissions and a parent role.

Update an existing authorization role by ID.

Delete an authorization role by ID.

List role assignments, optionally filtered by user, role, or domain.

Assign a role to a user, optionally scoped to a domain.

Delete a role assignment from a user by its assignment ID.

List authorization teams, optionally filtered by search term or active flag.

Get a single authorization team by ID.

Create an authorization team within an administrative domain.

Update an existing authorization team by ID.

Delete an authorization team by ID.

List members of an authorization team with pagination.

Add a user to an authorization team.

Remove a user from an authorization team.

List resource shares, optionally filtered by resource, target, or scope.

Get a single resource share by ID.

Share a resource with a scope and optional target and permission level.

Update the permission level of a resource share.

Delete a resource share by ID.

Query the authorization audit log with filters and pagination.

Get the calling user effective permissions for a set of resources.

Create a memory base for a flow (experimental, hidden in Langflow OpenAPI).

List memory bases, optionally filtered by flow with pagination.

Get a single memory base by ID.

List sessions stored in a memory base with pagination.

List messages stored in a memory base, optionally filtered by session.

Update an existing memory base by ID.

Delete a memory base and its stored memories by ID.

Flush a memory base session, persisting captured memories.

Check a memory base for configuration or embedding mismatches.

Regenerate a memory base, rebuilding its stored memories.

Test a connection to a knowledge base backend.

List available knowledge base ingestion connectors.

Ingest a local folder into a knowledge base with chunking options.

Ingest content into a knowledge base from a configured connector source.

Get the available metadata keys for a knowledge base.

List ingestion runs for a knowledge base with pagination.

Get a single knowledge base ingestion run by ID.

Reload an extension bundle (requires LANGFLOW_ENABLE_EXTENSION_RELOAD on the server).

Get extension events for the current user, optionally since a UTC epoch cursor.

Get an agentic workspace file, optionally as a downloadable payload.

Reset an agentic session, optionally targeting a specific session ID.

Get localized note translations for a flow by ID.

Get job queue metrics (superuser-only).

List available A2A (Agent-to-Agent) agents (Langflow 1.11.x). Requires server-side A2A enablement (LANGFLOW_A2A_ENABLED).

Get the A2A agent card (.well-known/agent-card.json) for a flow (Langflow 1.11.x). Requires server-side A2A enablement.

Invoke a flow via the A2A JSON-RPC endpoint (Langflow 1.11.x). The JSON-RPC envelope is a passthrough; validation lives server-side.

List pending human-in-the-loop requests for a required flow ID (Langflow 1.11.x).

Re-attach to a workflow job event stream by job ID (Langflow 1.11.x). Streaming is not incrementally surfaced; the initial payload is returned.

Resume a paused human-in-the-loop workflow with the pending request ID and optional decision (Langflow 1.11.x).

Run a public (unauthenticated-eligible) workflow (Langflow 1.11.x). Stream-only execution; streaming is not incrementally surfaced.

Get the Langflow Kubernetes-style readiness report from /healthz (Langflow 1.12.x).

List policy-visible model providers with stable IDs and display names (Langflow 1.12.x).

Read the install-wide approved model-provider policy (superuser-only Langflow 1.12.x API).

Replace the install-wide approved model-provider policy (superuser-only Langflow 1.12.x API).

Read the global blocked-component catalog policy (superuser-only Langflow 1.12.x API).

Replace the global blocked-component catalog policy (superuser-only Langflow 1.12.x API).

Read the global blocked-template catalog policy (superuser-only Langflow 1.12.x API).

Replace the global blocked-template catalog policy (superuser-only Langflow 1.12.x API).

Count flows using each catalog component (superuser-only Langflow 1.12.x API).

List flows affected by blocking a catalog component key (superuser-only Langflow 1.12.x API).

Read the atomic provider-and-catalog policy bundle (superuser-only Langflow 1.12.x API).

Atomically replace the provider-and-catalog policy bundle using an expected revision (Langflow 1.12.x).

List historical policy bundle revisions (superuser-only Langflow 1.12.x API).

Roll back the policy bundle to a prior revision with optimistic concurrency protection (Langflow 1.12.x).

Create or update a project at a caller-supplied ID using Langflow 1.12.x PUT upsert semantics.

Run the Langflow agentic assistant headlessly and persist its flow changes (Langflow 1.12.x SSE API).