Backlog

Next up

In Progress

Done

Main Roadmap

Some people have asked for this. Making a ticket to gauge the interest.

In Review

Add your own API key for various AI providers

alex@aiflowchat

A webhook node would greatly increase the connectivity of AI Flow Chat, enabling taking action on other apps and platforms, also receiving webhooks could be used as inputs for apps returning data by webhook.

Webhook functionality

Frederik Beier

Summary:Add the ability to draw and label visual groups directly on the canvas — for example, to mark a section of the workflow as “Editorial Phase” or “Psychology Review.”<div data-type="horizontalRule"><hr></div>Problem:When building complex flows with many interconnected nodes, it becomes difficult to understand at a glance which parts of the system belong to which stage of the process. For example, in a story-generation flow, several nodes might make up an Editorial Phase, but currently there’s no way to visually group or annotate them other than labelling the nodes individually. This makes it harder to explain the flow to collaborators, onboard new users, or even maintain the system over time.<div data-type="horizontalRule"><hr></div>Proposed Solution:Enable users to:<ol><li>Draw a bounding box or freeform shape around multiple nodes.</li><li>Assign a label (e.g. “Story Writing,” “Review,” “Production”).</li><li>Choose a background colour or outline style for quick visual grouping.</li><li>Optionally, collapse/expand these labeled areas for cleaner navigation.</li></ol><div data-type="horizontalRule"><hr></div>Why It Matters:<ul><li>Clarifies the logic and structure of complex flows.</li><li>Makes multi-phase workflows (e.g. story creation, data pipelines, product design) easier to communicate and present.</li><li>Reduces cognitive load by letting users treat clusters of nodes as single conceptual units.</li><li>Supports professional use cases like demos, audits, and investor presentations.</li></ul><div data-type="horizontalRule"><hr></div>Example Use Case:In a Story Creation flow:<ul><li>“Story Idea Generation”</li><li>“Editorial Review”</li><li>“Final Assembly &amp; Publishing”</li></ul>Each phase could be visually grouped and labeled for intuitive understanding and presentation.

Draw + Label Areas on the Flow Canvas

Dave Law

Core IdeaImplement a "Human In The Loop" or "Human Approval" node in workflows that requires human confirmation before continuing, ensuring high-quality outputs without unnecessary resource usage.<div data-type="horizontalRule"><hr></div>How It Works<ol><li>Integration: Place a "Human Approval" node at strategic points within a workflow.</li><li>Notification: Trigger notifications to relevant human operators for review.</li><li>Review: Provide a notification or interface that allows operators to view and evaluate content.</li><li>Approval: Allow operators to approve or reject the content.</li><li>Continuation: Once approved, the workflow proceeds to the next step.</li></ol><div data-type="horizontalRule"><hr></div>Key Features<ul><li>Content Display: Show generated content for review, such as images or text outlines.</li><li>Multi-step Approval: Enable multiple approval nodes in complex workflows for thorough vetting.</li><li>Notification System: Alert users when action is needed, reducing wait times.</li></ul><div data-type="horizontalRule"><hr></div>Main Benefits<ul><li>Quality Assurance: Enhances the accuracy and relevance of outputs through human oversight.</li><li>Resource Efficiency: Minimizes waste by halting workflows when early outputs are of subpar quality.</li><li>Flexibility: Adaptable to various use cases, especially in creative and detailed tasks.</li><li>User Empowerment: Enables users to make informed decisions, enhancing workflow outcomes.</li><li>Scalability: Supports complex processes through multiple approval layers.</li></ul><div data-type="horizontalRule"><hr></div>Incorporating a "Human In The Loop" approval node enhances workflow quality control, prevents resource wastage, and ensures that only approved content advances. This feature is crucial for maintaining high standards in workflows that rely heavily on creative judgment and precision.

Human-In-The-Loop / Approval Node

Have a shared context space where you can put various information and media in and share it across many flows and chats - but managed globally.If you change any information in the ressources folder, it would affect the places where it is referenced.Use cases it enables:<ul><li>One folder for all your different client’s material, e.g. videos, pdfs, notes, etc</li><li>One folder for a specific format: “viral hooks”, “viral thumbnails” etc</li></ul>

Global Ressources Folder

<h3>Core Idea</h3>Introduce a "Voice/Speak" node functionality that allows users to generate MP3 files from provided text using text-to-speech services like ElevenLabs. This feature will cater to various needs such as creating audio versions of written content, enhancing accessibility, and facilitating multimedia content creation.<div data-type="horizontalRule"><hr></div><h3>How It Works</h3><ol><li>Input Text: Users provide text input that they want to convert into speech.</li><li>Choose Service: Select a text-to-speech service like ElevenLabs.</li><li>Generate Audio: The system processes the text through the selected service.</li><li>Output File: An MP3 file is generated and ready for download or integration.</li></ol><div data-type="horizontalRule"><hr></div><h3>Key Features</h3><ul><li>Service Integration: Seamless integration with leading text-to-speech services.</li><li>Format Options: Output in MP3 or other audio formats as required.</li><li>Custom Voices: Ability to choose from different voice profiles and accents.</li><li>Script Automation: Pair with YouTube Transcript feature for dynamic script generation.</li></ul><div data-type="horizontalRule"><hr></div><h3>Main Benefits</h3><ul><li>Accessibility: Convert articles and text content to audio, making it accessible for visually impaired users.</li><li>User Experience: Enhance user engagement with audio options for content consumption.</li><li>Content Creation: Streamline the creation of audio content for podcasts, videos, and more.</li><li>Productivity: Automate the transformation of scripts from transcripts to audio, saving time in content production workflows.</li></ul><div data-type="horizontalRule"><hr></div>The Voice/Speak node is a versatile and efficient solution that addresses varying content creation and accessibility needs. By integrating it alongside features like the YouTube Transcript, it empowers users to create engaging audio content effortlessly, thus enhancing both reach and user experience.

Voice/Speak Node for Audio Generation

adding integrations like google drive would be great to drop7get files straight from a cloud drive.

Google drive integrations

lucaup

It would be great to have this. That would save a ton of time, not having to create pdf etc to be able to uplead info

CSV integration

Adam Haluska

Hello! I try to have a couple of group nodes to feed information to a chat (one is my videos I’ve recorded to learn my talking style, another is a reference video of another creator). But the chat can’t tell which video is mine vs. the reference (even though I put a label on the group). Is there a way around this/fix this?

Chat's Ability To See Labels

Joe Siddoway

I’ve had this happen multiple times where I create a new flow, add some text, create some groupings, etc. Then I exit and some of my items just disappear. Not all of them but some. Definitely doesn’t invoke confidence when I keep having things disappear on me

Keep losing data when exiting a flow

Trevor Bentley

<h2>Core Idea</h2>Introduce a “Database/Table View” feature for apps that allows users to input data in a tabular format, execute tasks in individual rows via a "Run" column, and manage the execution of selected or all tasks seamlessly. This facilitates organized task management, bulk data processing, and effective data imports/exports.<div data-type="horizontalRule"><hr></div><h2>How It Works</h2><ul><li>Table Layout: Users can input data into a table-like structure where each row represents a task or entry awaiting execution.</li><li>Run Column: Each row includes a "Run" button, enabling task execution upon completion of necessary inputs.</li><li>Run All Options: Users can choose to "Run all" tasks, or "Run all selected" for specific rows, ensuring control over which tasks to execute.</li><li>CSV Integration: Provide options for CSV import/export to facilitate bulk data handling and integrations with other systems.</li></ul><div data-type="horizontalRule"><hr></div><h2>Key Features</h2><ul><li>Data Entry Table: A spreadsheet-like interface for entering and managing input data.</li><li>Individual Execution: A dedicated column for executing each task, ensuring user control and flexibility.</li><li>Bulk Execution Options: Options to execute all tasks or only selected ones, enhancing efficiency.</li><li>CSV Input/Output: Support for CSV import/export for easy data integration and automation possibilities.</li></ul><div data-type="horizontalRule"><hr></div><h2>Main Benefits</h2><ul><li>Efficiency: Streamlined process for managing and executing tasks individually or in bulk.</li><li>Flexibility: Control over which tasks to run, allowing for selective execution.</li><li>Scalability: Ability to handle large volumes of data through CSV support.</li><li>Automation: Enhances the potential for integrating with external systems for seamless data transfer and automation.</li></ul><div data-type="horizontalRule"><hr></div>Implementing a "Database/Table View" feature offers users a powerful way to manage and execute tasks within an application, while also supporting automation through CSV integration. This feature meets user needs for organized, efficient, and scalable task management and execution.

Planned

Database/Table View for apps

Implementing a fact-checking process to highlight content requiring verification before publication.Possible services that I know of<ul><li><a target="_blank" rel="noopener noreferrer nofollow" class="link" href="https://www.factiverse.ai/api">https://www.factiverse.ai/api</a></li></ul>

Fact-checker node

Users want to be able to insert previous prompts into the text in the nodes. E.g. {{ node_label }} or something.

Make text input as variables inside prompt nodes

Having a HTTP POST block would open doors to access data via external APIs. The output, in JSON, can be temporarily post-processed by a fast tier prompt block, until further enhancements are available. <img data-width="100%" data-align="left" data-featurebase-content-key="680d0f796e266417058b91d2/post/685d5e9aa19da13adaa9f86a/0197acb6-593a-7388-9c54-50c136ba89c7/b64u-U2NyZWVuc2hvdCAyMDI1LTA2LTI2IGF0IDguMTguMDbDosKAwq9QTS5wbmc.png" src="https://680d0f796e266417058b91d2.featurebase-attachments.com/c/post/685d5e9aa19da13adaa9f86a/0197acb6-593a-7388-9c54-50c136ba89c7/b64u-U2NyZWVuc2hvdCAyMDI1LTA2LTI2IGF0IDguMTguMDbDosKAwq9QTS5wbmc.png?X-Amz-Expires=3600&amp;X-Amz-Date=20261001T040000Z&amp;X-Amz-Algorithm=AWS4-HMAC-SHA256&amp;X-Amz-Credential=DO801TYC4FCVNNEKURKM%2F20261001%2Ffra1%2Fs3%2Faws4_request&amp;X-Amz-SignedHeaders=host&amp;X-Amz-Signature=8b7cc75984d2fa0d28cb4418797387f64828e690d945d65cc620999deada3d5c" data-featurebase-content-filename="Screenshot 2025-06-26 at 8.18.06â¯PM.png">Open to any further suggestions to this feature enhancement.

HTTP POST Request Block

I would like a feature to insert files as resources that can be referenced in prompts. These files and resources may include brand tone-of-voice guidelines, internal knowledge, and other information that the AI might need to craft appropriate responses based on my prompts.

Ressources block/node

If a specific output format is needed, a formatter node could be included for the most commonly used formats.

Formatter node

Specify if input is optional or mandatory when running an app

When I copy nodes and resources from one board to another, they are pasted and looks great, but then when I close the board, they disappear like they were never there.Only natively uploaded assets remains within the board. The copy pasted ones all disappear.<img data-width="100%" data-align="left" data-featurebase-content-key="680d0f796e266417058b91d2/post/68d2a372e2c132a2dc42b2f8/019976cd-ebd7-77d8-ab69-b838ff035d7b/b64u-U2NyZWVuc2hvdCAyMDI1LTA5LTIzIGF0IDE1LjI2LjU5LnBuZw.png" data-featurebase-content-filename="Screenshot 2025-09-23 at 15.26.59.png" src="https://680d0f796e266417058b91d2.featurebase-attachments.com/c/post/68d2a372e2c132a2dc42b2f8/019976cd-ebd7-77d8-ab69-b838ff035d7b/b64u-U2NyZWVuc2hvdCAyMDI1LTA5LTIzIGF0IDE1LjI2LjU5LnBuZw.png?X-Amz-Expires=3600&amp;X-Amz-Date=20261001T040000Z&amp;X-Amz-Algorithm=AWS4-HMAC-SHA256&amp;X-Amz-Credential=DO801TYC4FCVNNEKURKM%2F20261001%2Ffra1%2Fs3%2Faws4_request&amp;X-Amz-SignedHeaders=host&amp;X-Amz-Signature=3d4207412333277a5b356696951c22fcf24786edb24342616a9b7c6fb80b6e5d">

Node Copy and Paste Bug

If you give a model too many images and the size of images exceeds 35MB in size, then most AI models can’t work with it

Request Entity Too Large: AI models can't handle more than 35MB for images

<ol><li>Multiple asset selection When I try to select multiple assets, I’m unsure when and hot to make it work. Sometimes work, sometimes doesn’t. I have not found the logic yet.</li><li>Multiple asset movement into groupsMoving items into the groups, doesn’t work. even tho I highlight and select multiple, it only pasts one inside. the rest remains outside. I have to move them one by one. </li><li>Group name label, constantly disappears. The group I’ve tried to move the items into, I have named 5 times already. It always disappears, after every new open / refresh. <video data-width="100%" data-align="left" data-featurebase-content-filename="Screen Recording 2025-09-26 at 16.37.55.mov" data-featurebase-content-key="eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJvYmplY3RLZXkiOiIwMTk5ODY3ZC0wZGIwLTdkNzktOTQ5ZS0zMWNiNTUyYjg3NjIiLCJ0eXBlIjoidmlkZW8tdXJsIiwicmVzb3VyY2VJZCI6IjY4ZDZhNzY3MGE4ZWI1NzYzNDJjM2FmYiIsInJlc291cmNlVHlwZSI6InBvc3QiLCJvcmdJZCI6IjY4MGQwZjc5NmUyNjY0MTcwNThiOTFkMiIsImlhdCI6MTc1ODg5ODAyM30.9zroWaF1RmRcjRVmzoMj4PhUf1rG_mSeIqFC2kUB10U" src="https://680d0f796e266417058b91d2.featurebase-attachments.com/c/videos/eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJvYmplY3RLZXkiOiIwMTk5ODY3ZC0wZGIwLTdkNzktOTQ5ZS0zMWNiNTUyYjg3NjIiLCJleHAiOjE3OTA4MzA4MDAsInR5cGUiOiJ2aWRlby11cmwiLCJvcmdJZCI6IjY4MGQwZjc5NmUyNjY0MTcwNThiOTFkMiIsInJlc291cmNlSWQiOiI2OGQ2YTc2NzBhOGViNTc2MzQyYzNhZmIiLCJyZXNvdXJjZVR5cGUiOiJwb3N0In0.Ox6xL7YakyHoIpr3zS2NhABNComj35VOzmRK5-4ZhtU/manifest.mpd"></video></li></ol>

Selecting and Moving multiple assets inside a group.

Integrate with more LLM providers, making the app model agnostic and being able to pick the best model for each prompt.Suggested providers<ul><li>Gemini</li><li>Llama</li><li>Claude</li><li>Grok</li><li>Mistral</li><li>DeepSeek</li></ul>

Making the flows model-agnostic

- Public (All) - Private (With the link) - Invite only (Registered users with invited emails)

Different access levels to publishing apps

I’ll see you guys around! <img data-width="25%" data-align="left" data-featurebase-content-key="680d0f796e266417058b91d2/post/6812958890f9cf7908e17f5f/01968897-cc3f-7a47-a553-038104a5e7ac/b64u-SU1HXzExNjAgMS5qcGc.jpeg" src="https://680d0f796e266417058b91d2.featurebase-attachments.com/c/post/6812958890f9cf7908e17f5f/01968897-cc3f-7a47-a553-038104a5e7ac/b64u-SU1HXzExNjAgMS5qcGc.jpeg?X-Amz-Expires=3600&amp;X-Amz-Date=20261001T040000Z&amp;X-Amz-Algorithm=AWS4-HMAC-SHA256&amp;X-Amz-Credential=DO801TYC4FCVNNEKURKM%2F20261001%2Ffra1%2Fs3%2Faws4_request&amp;X-Amz-SignedHeaders=host&amp;X-Amz-Signature=5c726ca34e51c70f7496319146d7237a8b47d04a5de3f9d2ecb74039161c35df" data-featurebase-content-filename="IMG_1160 1.jpg">- Alex

Completed

Hi there! Use this place to report bugs or request features!

I’ve setup a number of boards, gotten them laid out how I want, but when I leave and come back, all the nodes are rearranged around the board. They’re still linked, but just get shuffled around so I have to re-move them. Even after fixing them, once leaving and coming back they get shuffled again.

Board keeps rearranging when leaving and coming back

Hi,  I want to report a recurring issue I’ve been experiencing with the Claude Sonnet 4.5 model.  Every time I try to generate a full YouTube script, the model stops midway and never completes the script, even though I clearly instruct it to continue until the end. I repeated this more than six times, and the model made the exact same mistake in every attempt.  Because of this bug, a large number of my credits were consumed without delivering the result I paid for. I’m attaching two screenshot examples, but I had many more failed attempts.  Please investigate this issue and fix the bug. Also, I would like my credits to be fully reset, since the failed generations were not my fault and I didn’t receive the expected output.  Thank you.

Tarik

I would like to upload files as knowledge, such as PDFs, documents, etc. Additionally, I would like the option to include links to publicly shared Google Docs.Having the ability to reference specific and up-to-date company policies would be an excellent way to ensure that the AI's output remains relevant as corporate guidelines change.

Files node

Error when running node during prompt run

Error when running node

Pewter Crunchy

I’m running calls with Claude Sonnet 4.5 - and it is running through to about 2,300 words and then stopping mid-sentence.

Stop random capping out of word or token count

marshwah

Please see the before and after photos. It says no image uploaded.I tried to rerun. didn’t work.So i need to reupload again. <img data-width="100%" data-align="left" data-featurebase-content-key="680d0f796e266417058b91d2/post/68d6ae27aba4303cc50f1933/01998697-91f7-736f-a7db-1ed4589fb912/b64u-YWZ0ZXIucG5n.png" data-featurebase-content-filename="after.png" src="https://680d0f796e266417058b91d2.featurebase-attachments.com/c/post/68d6ae27aba4303cc50f1933/01998697-91f7-736f-a7db-1ed4589fb912/b64u-YWZ0ZXIucG5n.png?X-Amz-Expires=3600&amp;X-Amz-Date=20261001T040000Z&amp;X-Amz-Algorithm=AWS4-HMAC-SHA256&amp;X-Amz-Credential=DO801TYC4FCVNNEKURKM%2F20261001%2Ffra1%2Fs3%2Faws4_request&amp;X-Amz-SignedHeaders=host&amp;X-Amz-Signature=4747c533afbfee6edce3a471898a8fb554c797690e1b320b99c9e19c8c6d34b5"><img data-width="100%" data-align="left" data-featurebase-content-key="680d0f796e266417058b91d2/post/68d6ae27aba4303cc50f1933/01998697-7e04-753c-ae4d-d88014f13cab/b64u-YmVmb3JlLnBuZw.png" data-featurebase-content-filename="before.png" src="https://680d0f796e266417058b91d2.featurebase-attachments.com/c/post/68d6ae27aba4303cc50f1933/01998697-7e04-753c-ae4d-d88014f13cab/b64u-YmVmb3JlLnBuZw.png?X-Amz-Expires=3600&amp;X-Amz-Date=20261001T040000Z&amp;X-Amz-Algorithm=AWS4-HMAC-SHA256&amp;X-Amz-Credential=DO801TYC4FCVNNEKURKM%2F20261001%2Ffra1%2Fs3%2Faws4_request&amp;X-Amz-SignedHeaders=host&amp;X-Amz-Signature=651a8fef171f7d97ea89f83a5ca07756a1efa9b11b94e0576bc02df2dbe8a226">

All my images disappeared

Hey dont know if we already have it planned but would be cool to have like a folder for all resources and group them together - and then can be linked to a prompt/chatKind of like similar to <a target="_blank" rel="noopener noreferrer nofollow" class="link" href="http://poppy.ai">poppy.ai</a> where you can input many resources in folder and have one output

Folder Group

Mairo Miguele Baterzal

i subscribed to the 30$ monthly plan 7 days ago and now I am back to FREE plan for no reason.

Subscription Gone and I need to USE AI flow chat

Sinan Abou Hamdan

Core IdeaEnable users to remove the "AI Flow Chat" logo and branding from embedded apps to better integrate with internal and external tools.<div data-type="horizontalRule"><hr></div>How It Works<ol><li>Toggle Option: Introduce a setting toggle within the app's configuration panel to enable or disable the display of branding elements.</li><li>Custom Branding: Optionally allow users to replace "AI Flow Chat" branding with their own, enhancing customization.</li></ol><div data-type="horizontalRule"><hr></div>Key Features<ul><li>Branding Toggle: Simple on/off switch to hide/expose branding.</li><li>Custom Logo Upload: Capacity to upload a custom logo or text replacement.</li></ul><div data-type="horizontalRule"><hr></div>Main Benefits<ul><li>Enhanced Integration: Seamlessly align the AI tool with company branding to maintain a consistent look across platforms.</li><li>Professional Appearance: Improve the perceived value of the tools when facing clients or external stakeholders.</li><li>Market Differentiation: Offer this as a premium feature to justify higher pricing tiers.</li><li>Control and Flexibility: Give users autonomy over their app presentation, tailoring it to specific needs.</li></ul><div data-type="horizontalRule"><hr></div>Providing a toggle to hide or customize the "AI Flow Chat" branding offers flexibility, professional integration, and a more cohesive user experience. This feature not only enhances the system’s value for end-users but can also serve as a differentiator in the market, supporting a tiered pricing structure.

Ability to hide the "AI Flow Chat" logo and branding from embeds

We nid the Image file / Image container as an input field for apps: “Upload image” + “Image from URL” and “Paste from Clipboard” would be nice.