Text-Out-Loud

Integrating AI tools in one chat platform

About TOL

Text-Out-Loud is a platform that integrates all the AI tools that you are registered in, in one chat platform. Text out loud comes from the term "think out loud" which is used during user interviews to encourage the user to tell what they are thinking about or go through their minds loudly. In TOL the word think is replaced with text, as the user can type whatever they imagine, think or want, TOL analyzes the inserted prompt, detects what needed to be generated (Code, video, image,...) and routes the prompt to the appropriate AI tool from the integrated tools list.

‍

Problem

If you want to use several AI tools, you need to go back and forward between multiple browser tabs. The second thing is that, if you have an imagination in mind or a wild idea that you are still not sure about, you need to do a mental effort on recalling and thinking which AI tool from those you are registered in, would be the best fit for fulfilling what you are imagining of. Also if you know exactly which tools to use, you might want to efficiently compare between the responses generated from multiple AI tools using the same prompt, to know which one provided or nearly provided what was on your mind, if so you should open these in comparison tools together and keep copy, pasting, generating, comparing,... ffffffff, this task would consume much time and effort, especially when it becomes a daily task.

Solution

An AI platform that aggregates all the AI tools that you use or registered in together, using the chat pattern, which is a commonly used and popular pattern that people use everyday. What you need is just texting out what is in your mind in a plain English prompt, TOL will then analyze your words, calling out the tool(s) that would best help you meeting your imagination, while facilitating the comparing operation by just one click to view how the multiple tools responded to your prompt and what they generated for you.

The challenge

The user control and freedom to choose whether they need to interact with a specific AI tool or they already know the exact type of response they need (web, video, sound, ...etc) and want to select it, or they just have a wild rough idea in mind and don't know how or in which format it should be generated. The last one was the most challenging.

Switching between the tools that were in charge of generating the response to view their responses contextually, in the same view and without obstructing the chat experience.

Applying a futuristic, modern style in the design and the micro-interactions, similar to what all the AI tools incubate nowadays.

Discoverability

Ensuring that all the options that are important and might be used for the task at hand are discoverable whenever needed in a contextual way. This is to avoid the choice paralysis that me occur when showing the user all the options at once.

AI Futuristic Style

Applying the AI futuristic style, that appears to be the coming design trend, in a way doesn't affect the accessibility nor the readability to keep the platform usable with no issues.

Exploration

With the rise of the generative AI everywhere and the appearance of new AI tools every day. Studying, using and trying these tools, was the drive for TOL. As there are a lot of tools that generate the same format or type (for example: code, image, audio, video, text, ...) although the quality of the response depends on the inserted prompt itself and how you describe it, according to the prompt engineering, but the AI model itself also matters. It seems that from now on the AI will be a copilot in our tasks, so the ease of access to them all is crucial.
‍

‍

My browser tabs while exploring tools

Desktop Research

People are asking for the best tool to satisfy their needs and solve their problems. All that they need is just drop in what they want or imagine of and sit back to see the results, but they don't know from where they could start, also they don't know which is the best tool that would generate the best response for them (comparison), so they go to the AI communities searching for answers to their questions (What is the best..., which tool could..., Is there any tool that..., what's better A or B). They are asking for others' experiences that could be a biased experience or simply don't meet their needs or personal preferences.

Facebook posts from multiple AI groups

Competitors

Pros and Cons of. Poe (Powered by Quora) which the most popular tool similar to TOL nowadays

‍

Screenshot from Poe web app


‍Pros
-
Has a mobile application in addition to the web app.‍

- Giving access to the paid tools like ChatGPT-4, with one prompt per day

- Automatically adding new tools to the list as Poe integrates with them, without doing any effort.


Cons
- Must firstly select the tool before inserting any prompts.

- Only the AI tools that Poe is integrating with can be used, no opportunity to add more.

‍

‍

‍

‍

‍

‍

The solution

Text-Out-Loud (TOL)

A platform that aggregates your AI tools in one place in a chat view.

‍

‍

‍

‍

Main View

A fixed side menu consisting of two sections, your chats and your integrated AI tools grouped together separately . The main focal area contains the primary content of the two way dialogue/chat, of the prompt provided to TOL and the response generated from whatever AI tool(s) that was in charge for generating the content.

‍

‍

‍

‍

Personalization Layer

TOL adds a personalization layer by creating interactive AI characters and giving them a detailed persona through backstory, habits, occupation, example of how the character should respond to a sample of questions and much more details using Mycharacter.ai, plus giving a visual representation for this character through a 3D model using Lumalabs.ai, so TOL merged both tools to generate an interactive, personalized and realistic character visualized in a 3D model.

‍

My Characters Expandable Drawer

‍

‍

Character Modal

‍

Watch the video for the prototype of selecting a character

‍


Generating a prompt while the type selected upfront

Users could be already aware of what they need, whether it's a code, video, image, audio, ..etc. So, they can choose the type of the response they need before typing in their prompt from the command menu that opens by typing "/" in the message field. This list of response types is auto-generated and updated from adding tools to your integrated tools list, TOL detects the response type that each registered tool can provide and appends it to the list. It also shows which tools are in each type beside it. After selecting the response type, the user can deselect one of the tools to stop it from generating a response

‍

Command menu

‍

‍

Response type selected

‍

‍

Prompt inserted & one tool (the middle one) is deselected

‍

‍

AI tools loading

‍

‍

Response with video type generated

‍

‍

Named entity recognition (NER)

As many users may not know what is the best tool that fits their needs or they may be unsure of what they exactly need, they just need to drop their imagination in plain language and TOL does the rest of the operation. TOL can detect and analyse the inserted prompt to know exactly what the user needs or imagining of, then directs the prompt to the appropriate AI tool from the integrated AI tools list. This is the main edge of TOL that makes it stand out from its competitors.

‍

User can type-in any message in plain language in the message field



And TOL will do the rest from detecting what is on the user's mind, what is the response format would be the best fit for their imagination, to suggesting the best tools for generating the response

‍
‍
Until the desired prompt is lastly generated
‍

Impact Overview

TOL should make people's lives better, accelerating their daily operations while saving effort and time consumed in tasks that could be automated through AI.

Saving time and effort

TOL reduces the time and effort consumed in the operation of selecting the tool, deciding on the needed prompt type and comparing between the response generated from multiple AI tools.

One bucket for all

Aggregating all the most used AI tools that the user is registred in, all in one platform designed in a chat pattern for the usability and ease of use of this popular and familiar pattern