This is a submission for the Agent.ai Challenge: Full-Stack Agent (See Details)
Overview
It all started with a simple conversation with my non-technical business partner:
"You know... I get random ideas and thoughts at the most unexpected times during the day. I wish I could just send them as voice notes and later use them to write my blog posts. But it's tough to find the time to sit down and go through all of them."
This sparked an idea—to create an AI-powered writing assistant that captures these spontaneous thoughts and transforms them into structured content, helping busy professionals stay consistent with their content creation efforts.
What I Built
I developed a full-stack multi- AI agent that seamlessly integrates with Telegram and Agent.ai to capture, process, and draft LinkedIn posts based on voice notes.
Why It Matters
One of the things I love about this solution is how incredibly convenient it is. I genuinely see myself using it over the next few months because of its effortless accessibility. With just a simple voice note, I can capture my thoughts anytime, anywhere, and have an AI agent ready to process them—whether I’m commuting, taking a walk, or in between meetings.
The best part? No need to log into any extra dashboards or complicated tools. Although Agent.ai’s dashboards provide valuable insights and controls, the ability to manage everything directly from my phone via Telegram makes it a game-changer for productivity on the go.
Voice Note Capture via Telegram Chatbot:
We understand that the best ideas can strike at any moment. That’s why we’ve built a tool that lets users capture them effortlessly through a Telegram bot. Simply by sending a voice note, and it will be transcribed and securely stored as an “idea” for future content. Behind the scenes, a webhook triggers an Agent.ai process to analyze and categorize the user's thoughts, ensuring they’re ready when they need them.
AI-Powered Writing Profile:
Your unique voice matters, and our AI ensures it shines through. Brain Dump continuously refines your writing style by analyzing your past LinkedIn posts and transcribed voice notes. It learns your tone, favorite phrases, and writing patterns to generate content that truly reflects your personality and professional brand.
Automated LinkedIn Post Drafting:
When the user is ready to post, they can request the AI agent to generate a draft based on stored ideas, again, all through a simple text or voice message.
The agent will then go through the user's profile and previous ideas to produce a well-structured draft using the user's tone of voice, preferred topics, and content style.
Demo
To bring Brain Dump to life, I utilized a set of specialized AI Agents, each playing a crucial role in processing and generating content seamlessly. Below are the key agents used in the system, along with their respective functions:
Orchestrator Agent
The brain behind Brain Dump, this agent is responsible for processing all incoming inputs from the Telegram chatbot. It analyzes whether the message is a new idea, a request for a summary, or an instruction to generate a post, and then routes it to the appropriate AI agent for further action.
Summarizer Agent
This agent provides users with concise and meaningful summaries of their stored ideas. Whether they need a quick recap of recent thoughts or an overview of specific topics, the Summarizer Agent extracts key insights to keep their content organized and accessible.
Linkedin Master
A specialized AI agent that focuses on analyzing the user’s LinkedIn activity, including past posts and writing patterns. It helps refine the user's writing profile, ensuring that generated content matches their tone, style, and frequently used phrases.
Simple LLM Fallback
Acting as a safety net, this agent ensures smooth handling of unexpected queries or when other agents are unable to provide a precise response. It helps maintain a seamless user experience by offering relevant suggestions and fallback responses when needed.
Telegram Chatbot API Gateway
In addition to the AI Agents, I developed a Telegram chatbot, which serves as the main interface for users to interact with Brain Dump. It acts as an API Gateway, forwarding user prompts—whether as text or voice notes—directly to the AI agents for processing.
You can test the implementation here 👉: t.me/YourBrainDumpBot
You can explore the chatbot’s implementation here:
👉 GitHub Repository
JoseCToscano
/
braindump
Simple Telegram Chatbot to send random voice notes with ideas for LinkedIn Posts
🚀 Brain Dump Telegram Chatbot
Brain Dump is a Telegram chatbot designed to help users capture their ideas effortlessly and turn them into engaging LinkedIn posts with the power of AI. Whether you’re on the go or brainstorming at your desk, simply send a voice note or text message, and let the AI do the rest!
🛠️ How It Works
-
Capture Ideas:
- Send a voice note or text message to the chatbot via Telegram.
- Your input is forwarded to an AI Orchestrator for processing.
-
Processing & Categorization:
- The orchestrator determines whether the input is
- A new idea to be stored.
- A request to summarize existing ideas.
- A command to generate a LinkedIn post.
- It routes the request to the appropriate AI Agent for further action.
- The orchestrator determines whether the input is
-
AI-Driven Insights:
- Specialized AI agents analyze your past LinkedIn posts and voice notes to refine your writing style.
- When you're ready, Brain Dump generates a…
How It Works
Idea Capture:
The user sends a voice note via Telegram.
Processing & Orchestration:
The Telegram bot forwards the message to an AI Orchestrator via a webhook.
The orchestrator analyzes the input and determines the appropriate next step.
Specialized AI Agents:
Depending on the input, different AI agents are triggered to:
Process and categorize new ideas.
Analyze existing ideas for insights.
Generate an initial draft if enough content has been collected.
Continuous Learning:
The AI refines the user's writing profile by analyzing voice notes and past LinkedIn posts, ensuring content stays aligned with their style.
Content Generation:
When the user requests a post, the AI drafts it based on their tone, writing patterns, and key themes.
Agent.ai Experience
I wasn't quite sure where to start, so I decided to dive in from scratch and explore everything. To my surprise, building AI Agents was way easier than I expected. Based on my past experience, I assumed I’d have to deal with API keys, set up Python integrations, and rely on APIs or SDKs to connect with AI models. However, the Builder made it as simple as selecting the features I needed, prompting for user inputs, and seamlessly passing them to other actions. Calling AI models was incredibly straightforward.
Once I figured out my application's needs, it was just a matter of sitting down and building it. I opted to use multiple AI Agents instead of one large agent to simplify debugging and allow for easier feature expansion down the line.
Getting everything to work via webhooks did take some effort, and I encountered a few challenges along the way. However, the Agent.ai team was super responsive and supportive, helping me resolve issues quickly.
PRO TIP HERE: When setting up a webhook, start by testing it with minimal or no user inputs. Once you confirm it's working correctly, gradually add inputs one at a time, testing each step along the way. Initially, I tried adding everything at once and testing the trigger, but I encountered several issues that needed to be fixed on the server. Taking a step-by-step approach can help prevent unexpected errors and streamline the setup process.
One of the things I love most is that you don’t need to know how to code to build AI Agents. I’m confident my non-technical co-founder will be thrilled with how intuitive it is. That said, having some coding experience really opens up possibilities for creating awesome features like Brain Dump and integrating with various services.
Whats next?
As Brain Dump continues to evolve, the focus remains on refining its ability to capture authentic tone and maintain consistency with the user's unique professional branding. Striking the right balance between automation and user control has been key to ensuring a seamless experience. Looking ahead, future improvements include expanding to other platforms like Twitter and Medium, incorporating sentiment analysis to better capture emotions, and providing context-aware recommendations based on industry trends.
Ultimately, Brain Dump empowers busy entrepreneurs and professionals by turning spontaneous ideas into engaging content effortlessly, helping them stay consistent and boost their online presence.
Top comments (1)
Such a great idea!! Lov it !