Flow Image and Video Workflows, Automated
AutoNano runs Google Flow from a compact Chrome side panel: images, videos, edits, batches, history, downloads, and MCP automation in one connected workspace.
AutoNano live workflow demo
Image
Single or batch prompts
Video
Text, keyframes, references
History
Sessions, previews, downloads
MCP
Claude, Codex, Cursor relay
Control surface
Images, videos, edits, history, settings, and MCP stay in one Flow cockpit.
6 tabs = 1 pipeline
The Problem
Flow Workflows Break When Setup, Media, and Automation Split Apart
AutoNano is designed around the real Flow loop: connect a project, create images or videos, edit sources, batch prompts, download media, keep history, and let external AI tools call the same pipeline.
Waiting inside Google Flow while repeating the same prompt setup by hand
Downloading images and videos without a clean session trail
Losing which model, ratio, frame, or reference produced the result
Refreshing auth, project context, and output settings across sessions
Running prompt lists manually instead of pausing, resuming, and tracking a queue
No bridge for Claude, Codex, or Cursor to generate through your Flow tab
AutoNano fixes all of this
One Flow project -> image, video, edit, batch, MCP
Prompts, media inputs, model fallback, downloads, and history stay attached to the session.
Command Deck
One side panel that runs the whole Flow pipeline.
Instead of scattering prompts, frames, references, downloads, and automation across tools, AutoNano keeps the workflow inside one compact control room.
Active module
Image Generation: Single or Batch
Create images from one prompt or switch into batch mode for prompt lists. Choose Nano Banana 2, Nano Banana Pro, or Imagen 4, then control ratio, count, references, downloads, and queue speed.
Prompt slots
Focused
Media inputs
Images + refs
Automation
Flow tab
History
Session-based
What this unlocks
Workflows
Generate, Edit, Batch, and Automate Without Losing Flow Context
AutoNano covers the actual side-panel loop: connect a Flow project, run media jobs, recover the session, download outputs, keep local history, and expose the same workflow to MCP clients.
Generate Images One Prompt or One Queue at a Time
Write a prompt for a quick run, or paste a full prompt list for batch mode. AutoNano keeps model, ratio, count, references, downloads, and queue progress in the same surface.
Generate Videos from Text, Frames, or References
Choose Text, Start, Start+End, or Ref mode, then generate one video or run a prompt list with shared keyframes and reference images.
AI Image Editing Workflow
Upload a source image, write the edit instruction, add optional references, and send the source into Flow as a base image with the right aspect ratio.
Let Claude, Codex, and Cursor Call the Same Flow Pipeline
Generate an MCP URL, connect an external AI client, and watch task activity as the server relay sends image, video, or edit work to your extension.
How It Works
From Flow Project to Downloaded Media
Step 01
Connect AutoNano to Google Flow
Sign in with Google, open Google Flow, enter a project page, and keep the Flow tab open while AutoNano checks the connection.
Step 02
Start with a Prompt, Frame, or Source Image
Create from text, upload video keyframes, add reference images, upload an edit source, or paste a batch prompt list.
Step 03
Choose Flow Settings Once
Select model, aspect ratio, output count, references, auto-download, folder, server target, or video mode before the run.
Step 04
Run, Download, Review, Automate
Track queue status, download images or MP4s, review sessions in History, and let MCP clients trigger the same pipeline.
one connected loop
Flow auth, media inputs, downloads, and MCP stay together
Core Capabilities
Core Controls That Match the Way Flow Work Is Made
AutoNano keeps Flow connection, model selection, references, keyframes, batch queues, downloads, history, and MCP automation in one connected extension experience.
Flow Model Selection
Choose Nano Banana 2, Nano Banana Pro, or Imagen 4 for image generation and editing, with automatic fallback when a model quota is exhausted.
1-4 Outputs Per Prompt
Generate 1, 2, 3, or 4 images or videos from a prompt so you can compare directions without repeating the setup.
Reference Images and Keyframes
Upload up to 4 references for images and video Ref mode, or use start and end frames to guide video motion.
Video Modes
Use Text, Start, Start+End, or Ref mode with Fast or Lite Veo 3.1, landscape or portrait ratio, and optional 1080p upscale.
Batch Control
Import .txt or .csv prompt lists, set queue delay, then pause, resume, stop, reset, and inspect per-prompt status.
MCP Automation
Connect Claude Desktop, Claude Code, Cursor, or Codex through an MCP URL so external tools can trigger generation.
Results can be reviewed, edited, downloaded, or called from external AI tools
Explore Flow ControlsBefore vs After
What Gets Easier When Flow Becomes a Control Panel
Connect to the right Flow project
Keep checking tabs and project URLs yourself
Connection check, project detection, and recovery prompts
Generate images from repeated prompt lists
Paste and click through Flow one prompt at a time
Shared image batch queue with pause, resume, and stop
Create videos from text, keyframes, or references
Rebuild mode, frames, and output settings per run
Video single and batch modes with shared visual inputs
Let AI coding tools request media
Copy prompts between chat, terminal, and browser
MCP URL for Claude, Codex, Cursor, and external clients
Review and download generated media
Search your Downloads folder with no session context
Local History grouped by Generate, Video, Edit, and Batch
The result?
Flow auth, prompts, references, keyframes, downloads, history, and MCP activity stay in the same product.
Designed for Repeatable Flow Automation
The product is organized around the real Flow loop: connect, configure, generate, edit, batch, download, review history, and expose the same work to MCP clients.
Flow Session Awareness
Detects the Google Flow tab, checks for an active project page, refreshes auth, and warns when the Flow tab needs attention.
Persistent Defaults
Set default image model, aspect ratio, images per prompt, batch delay, output folder, auto-download, and server target.
Image and Video Inputs
Use image references, edit source images, video start frames, end frames, and reference images from one side-panel workflow.
Batch Visibility
Track prompt queues through pending, running, done, and failed states with pause, resume, stop, and reset controls.
Local Session History
Review generated images, edited outputs, videos, and batch sessions with previews and download actions.
MCP Activity Feed
Watch Claude, Codex, Cursor, and other MCP calls as they process, complete, fail, and return media previews.
Six Tabs That Cover the Real Flow Loop.
AutoNano connects Flow project detection, image and video generation, edit workflows, local history, MCP automation, and output settings in one Chrome Extension.
Built for repeatable Flow work
Defaults, History, and MCP activity help you keep context across sessions instead of starting from zero.
Default Controls
Model, ratio, count, delay
Output Folder
Choose where media lands
MCP URL
Connect external AI clients
Free Limits
5 image prompts, 3 video prompts, 20 history items
Pro is already designed into the product.
The current public build runs on the Free plan. Pro is planned for unlimited image batch, unlimited video batch, full generation history, and priority support.
$9
planned
Operational Backbone
The invisible parts are part of the product.
AutoNano is more than a prompt form. It manages auth, Flow recovery, local persistence, background automation, quota decisions, downloads, server connection, and MCP keys behind the side-panel UI.
Google sign-in + AutoNano JWT
AutoNano signs in with Chrome Identity, can refresh Google tokens silently, exchanges the Google token for an AutoNano JWT, and uses that JWT for profile, MCP keys, analytics, and the server relay.
Flow tab recovery
The extension checks for a Flow tab, verifies the active project page, refreshes auth from inside that tab, reloads after long idle periods, and resets the Flow tab after batch completion.
Background worker pipeline
A Manifest V3 service worker uploads media, starts image batches, runs edits, generates and polls videos, downloads files, checks credits, and emits status back to the side panel.
Quota and fallback handling
When an image model hits per-model daily quota, AutoNano tries the fallback chain: Nano Banana 2, Imagen 4, then Nano Banana Pro. If all fail, it checks Flow credits before stopping.
Local persistence
The side panel keeps signed-in UI state, active tab, sub-tabs, batch composer, and local History. Chrome storage keeps Flow settings, AutoNano JWT, and Google auth user data.
MCP keys and activity
Users can generate, copy, regenerate, or revoke an MCP URL. MCP calls appear in an activity feed with connector badges, processing state, previews, downloads, and failure details.
Server + relay
AutoNano talks to both Flow and the AutoNano API.
The Flow tab is used for authenticated generation. The AutoNano server is used for sign-in exchange, user profile, MCP keys, analytics, and the WebSocket relay that lets external tools dispatch work to the extension.
Settings coverage
Every default has a job.
Generation defaults
Output
Server target
Runtime messages
FAQ
Frequently Asked Questions
Turn Google Flow Into a Side-Panel Control Room
Connect a Flow project, run images and videos, edit source assets, batch prompt lists, download outputs, review local history, and let MCP clients call the same pipeline.