Miro · Miro AI Workflows

Flows

A user-initiated chain of Miro Formats and AI Instruction blocks on the canvas, in which each step's prompt reads connected content and prior output and generates persistent, editable Miro objects such as Docs, diagrams, tables, Kanbans, slides, prototypes, images and complete board layouts.

Recorded characteristics

Function
Miro documents Flows as chaining Formats on the canvas to build AI-powered workflows, where each Format acts as input for the next. A user adds a supported Format or AI Instruction block, connects it with diamond AI connectors, opens the TASK bar above the Format and writes a prompt that may use output from any connected Format or instruction block, then clicks Run. The Flow selected context menu shows how many steps remain to be executed, and the run can be started from any Format or block in the Flow. Supported Formats are Board Layout, Diagrams, Docs, Images, Embed iFrame Code, Kanban, Prototypes, Slides, Tables and Timeline. Miro documents a Board Layout generation node in a Flow: the user describes the layout, runs the Flow, and Miro AI generates a structured editable board containing frames, stickies, shapes, tables, Kanbans and interactive widgets. Miro documents that custom Flows can be saved and shared so anyone in the team can run the same workflow, and that readymade Flow templates can be used or customised.
Data access
Miro documents that "your canvas is the prompt": sticky notes, Docs, Tables and images already created on the board can become the starting point for a Flow, and a prompt can describe how the Format reads input content or any content on the board. Miro documents that the TASK box lets the user select a large-language model, a knowledge base or web search, with options varying by Format, and that Knowledge integrates providers including Amazon Q (Beta), Gemini Enterprise, Glean, Microsoft Copilot (Beta) and Miro Insights, plus web search, for retrieval into Formats such as Docs, Tables, Sticky notes and Slides.
Actions
Can take actions
External actions
Unknown
Human confirmation
Required
Permission basis
Mixed
Administrative control
Miro documents Enterprise Company admin AI capability controls in the Admin Console with Everyone, No one or Specific teams settings, including a dedicated Flows category described as "Connect AI prompts to canvas content to create automated workflows", plus separate Sidekicks, Prototyping, content creation, images and beta feature categories, with individual feature management available through the Enterprise Guard add-on. Miro documents that Company admins must enable Miro AI for an organisation and can enable it company-wide or for specific teams, and that Company admins must enable and approve each Knowledge integration. Miro documents billing and Company admin visibility of AI credit balances and usage analytics.
Default state
Conditional
Availability
Miro documents Flows for team members on Free, Starter, Education, Business and Enterprise plans, on browser, desktop and mobile. Miro documents that on Free, Starter and Education each team member has five Flow runs that do not reset, that the Flows public beta has ended and the maximum run count now applies, and that an indefinite number of runs depending on AI-credit allocation requires the Business plan. On Enterprise, Miro documents that an Accelerate license is required.
Licensing
Miro documents that Flows consume AI credits and that credits consumed per run depend on output type, giving one credit for a Sticky, five credits per AI-generated screen for Prototypes and five credits per board layout generation. Miro documents pooled monthly credit allocations per plan (Free 25/month, Starter 25/license/month, Business 50/license/month, Education 100/month, Enterprise varying by license tier), no rollover, purchase of additional credits by billing or company admins, and that AI actions pause when a team exhausts its credits. Miro documents that guests and visitors cannot run AI actions.
External model or provider
Miro documents Select Your Own Model (Beta) for Flows and Sidekicks, with a default model shown as Recommended. For Flows the model selector applies to Docs and Image Formats and to the Instruction Block, so different steps can use different models within a Flow. Miro lists reasoning models GPT-5, GPT-5-mini, GPT-5-nano, Claude Sonnet 4, Claude Sonnet 4.5, Gemini 2.5, Gemini 2.5 Flash, o3 and o4-mini, non-reasoning models GPT-4o, GPT-4o-mini, GPT-4.1, GPT-4.1-mini, Claude Sonnet 3.7, Claude Sonnet 3.5 and Gemini 2.5 Flash Lite, and image models Stable Diffusion 3.5 Large and Gemini 2.5 Flash Image (Nano Banana); an example default shown is AWS | Claude 3.5. Miro documents that models may be hosted by Miro or in Microsoft Azure AI.
Limitations and uncertainty
Every documented Flow run is started by a person clicking Run; Miro's Flows documentation establishes no schedule, event trigger, self-trigger or external trigger, and scheduling behaviour documented for the separate Miro Automations feature is not attributed here. Miro does not document what happens between steps once a run starts beyond a remaining-step counter, so per-step confirmation is not established. Miro documents a Revert control restoring any Format version from the last 24 hours and gives the example of accidentally running a Flow and overwriting a Doc, which establishes that Flow output can overwrite existing Format content; broader authority to edit arbitrary existing board objects is not established. External action capability is recorded as unknown: Knowledge providers and web search are documented for retrieval only, and Miro does not state whether Flows can write to external systems. Board-permission inheritance for Flow execution is not explicitly documented. Authority documented for Sidekicks, Create with AI, Action Shortcuts, Miro Automations, the Miro MCP Server and the Miro agent for Copilot is excluded from this record.

Evidence