DOCSEARCHbyAlgolia
Submit Docs

AI Flow Documentation

website·Indexed 33 days ago

https://docs.ai-flow.net

Documentation for AI Flow's node-based workflow system, featuring a comprehensive library of utility nodes, tool nodes for image, video, text, and JSON processing, web integration nodes (HTTP GET/POST, Google Search), and model nodes such as Stable Diffusion.

Tokens
33.6K
Snippets
102
Records
404
Agent score
54%

What's inside AI Flow

  1. Overview of ElevenLabs node capabilities

    The ElevenLabs node integrates ElevenLabs audio APIs into a flow. It requires a personal ElevenLabs API key stored in the Secure Store to function. The node supports four primary modes:

    • Create Speech: Converts text into audio using a specific voice.
    • Voice Changer: Transforms an uploaded audio file into a different ElevenLabs voice.
    • List Voices: Retrieves a list of available voices with optional filtering.
    • List Models: Retrieves a list of available ElevenLabs models for speech generation or conversion.

  2. Overview of Video Tools node capabilities

    The Video Tools node allows for various video manipulation tasks. The node reconfigures its available fields based on the selected mode. Supported modes include:

    • Trim Video: Cut video by time range (Output: Video)
    • Resize Video: Change video dimensions (Output: Video)
    • Merge Videos: Combine multiple videos (Output: Video)
    • Add Watermark: Add image overlay to video (Output: Video)
    • Extract Audio: Remove video and keep audio (Output: Audio)
    • Convert to GIF: Convert video to GIF image (Output: Image)
    • Convert to MP4: Convert video to MP4 format (Output: Video)
    • First Frame: Extract the first frame as an image (Output: Image)
    • Last Frame: Extract the last frame as an image (Output: Image)
    • Reverse Video: Play video in reverse (Output: Video)
    • Remove Audio: Strip audio track from video (Output: Video)

  3. Overview of AI-Flow Cloud hosted platform

    AI-Flow Cloud is the Pro hosted version of AI-Flow. It provides a managed environment where configurations, updates, and security patches are handled automatically, removing the need for manual setup or maintenance. All tools are available by default upon deployment.

  4. Overview of Airtable node capabilities

    The Airtable node allows users to interact with Airtable data through several operational modes. All modes output data in JSON format and require a valid Airtable access token, Base ID, and Table Name.

  5. Overview of Variable scopes and operations

    The Variables node is used to store and retrieve state within a flow. It supports three distinct scopes for data persistence:

    • Run: Temporary state that exists only for the duration of a single execution.
    • Flow: State that is persisted specifically for a saved flow.
    • Global: State that is persisted per user account across all flows.

    The node supports four operational modes: Get Variable, Set Variable, List Variables, and Delete Variable.

  6. Overview of AI-Flow visual workflow platform

    AI-Flow is a visual platform used to combine AI models, data, media, and external services into directed graph workflows. Instead of manual data movement, users connect reusable nodes where the output of one step serves as the input for the next. The platform supports providers such as OpenAI, Google Gemini, Anthropic, Replicate, fal.ai, Stability AI, and ElevenLabs.

  7. Overview of Runner MCP

    The Runner MCP enables an AI assistant to execute existing workflows in an AI-Flow account and read the results without requiring the user to open the AI-Flow interface. It uses a Streamable HTTP connection authenticated via an API key.
    Endpoint: https://mcp.ai-flow.net/runner

  8. Overview of the AI-Flow API Dashboard

    The API Dashboard is a dedicated operational view for monitoring workflows launched through the AI-Flow REST API. It allows developers to track completed, pending, or errored runs, search by run ID or workflow name, and compare execution durations. Note that this dashboard specifically tracks API-initiated runs; interactive runs performed within the workflow editor are managed separately.

  9. Overview of AI-Flow MCP Servers

    AI-Flow provides two complementary Model Context Protocol (MCP) servers that allow AI assistants to interact with workflows. The Builder is used to create and edit workflows, while the Runner is used to execute existing workflows. Both servers complement the visual editor, allowing users to switch between AI-driven interaction and manual editing.

  10. Overview of Builder MCP for workflow creation

    The Builder MCP is a tool currently in development designed to create AI-Flow workflows from plain language descriptions. Unlike the Runner MCP (which executes existing workflows), the Builder assembles the workflow graph, validates it, and saves it to the user's account for further refinement in the AI-Flow visual editor.

    Key capabilities include:

    • Intent-based node discovery: Finding nodes by what they do rather than their specific names.
    • Automated assembly: Adding nodes, wiring inputs/outputs, and setting field values.
    • Pre-save validation: Detecting cycles, type mismatches, and disconnected inputs before saving.
    • API Readiness: Automatically adding api-input and api-output nodes to make flows compatible with the Runner MCP.
    • Subflow reuse: Integrating existing logic pieces into new workflows.

    Constraints: The Builder cannot publish templates and cannot execute workflows (running is handled by the Runner MCP).

  11. Overview of Smart Node capabilities

    The Smart Node is a versatile AI tool that adapts its configuration and model selection based on the selected task. It simplifies the workflow by removing the need to manage multiple specialized nodes. Available modes include:

    • Generate Image: Create a new image from a prompt.
    • Edit Image: Modify an existing image or use images to generate a new one.
    • Upscale Image: Increase the quality/resolution of an image.
    • Caption Image: Generate a text description of an image.
    • Generate Video: Create a video from a prompt and optional audio.
    • Animate Image: Create an animation between a first and last frame image.

  12. Overview of JSON Tools node modes

    The JSON Tools node allows for the extraction and manipulation of JSON data. If the input JSON is embedded within text, the node automatically uses the first valid JSON object it finds. The node's available fields change based on the selected operational mode.