I’m John, founder of PostReach AI. For the last 12 months, my team has been building a conversational agent for social media.
The Beta: We are opening our beta today for the HN community. It is completely free while we gather feedback.
A Note on Access: > Because we are managing heavy GPU loads for the visual analysis and content generation, we are provisioning new accounts in batches. You’ll receive your access email within roughly 30 minutes of signing up. We are doing this to make sure the "Brand Brain" training for each new user doesn't crash the system for everyone else.
The Problem: As founders, we all know we should be active on LinkedIn, X, Facebook and Instagram. But the reality is a fragmented mess, you use ChatGPT for ideas, Canva for design, and Hootsuite for scheduling. It’s a context-switching tax that most small teams can’t afford.
What we built: PostReach isn’t a dashboard, it’s a chat interface. We wanted to see if we could collapse the entire "idea-to-published" stack into a single conversation.
Some technical bits the community might find interesting:
-Brand Ingestion: We don't just prompt a generic LLM. We built a "Brand Brain" that ingests your website and past voice to ensure the AI doesn't sound like a generic bot.
-Visual Analysis: Our Image-to-Post feature uses visual reasoning to understand the "mood" of an image. If you upload a photo of your messy dev desk, it won't just say "Working hard", it understands the hardware and context to write something relatable.
-URL-to-Post: We use a custom summarization pipeline to extract the "hook" from any article or blog post and transform it into platform-specific formats (a technical post for X vs. a professional takeaway for LinkedIn).
So why now? We’ve just moved into beta after aggressive testing. We are looking for our first "real world" cohort of users.
The Ask: We would love for the HN community to break this.
-Does the conversational flow feel faster than a dashboard?
-Is the AI output "cringe" or actually usable for a technical audience?
-What’s missing from your current workflow that we should automate next?
We’re around all day to answer questions about the tech stack, the 12-month dev cycle, or our thoughts on the future of autonomous agents in marketing.
John_V•1mo ago
I’m John, founder of PostReach AI. For the last 12 months, my team has been building a conversational agent for social media.
The Beta: We are opening our beta today for the HN community. It is completely free while we gather feedback.
A Note on Access: > Because we are managing heavy GPU loads for the visual analysis and content generation, we are provisioning new accounts in batches. You’ll receive your access email within roughly 30 minutes of signing up. We are doing this to make sure the "Brand Brain" training for each new user doesn't crash the system for everyone else.
The Problem: As founders, we all know we should be active on LinkedIn, X, Facebook and Instagram. But the reality is a fragmented mess, you use ChatGPT for ideas, Canva for design, and Hootsuite for scheduling. It’s a context-switching tax that most small teams can’t afford.
What we built: PostReach isn’t a dashboard, it’s a chat interface. We wanted to see if we could collapse the entire "idea-to-published" stack into a single conversation.
Some technical bits the community might find interesting:
-Brand Ingestion: We don't just prompt a generic LLM. We built a "Brand Brain" that ingests your website and past voice to ensure the AI doesn't sound like a generic bot.
-Visual Analysis: Our Image-to-Post feature uses visual reasoning to understand the "mood" of an image. If you upload a photo of your messy dev desk, it won't just say "Working hard", it understands the hardware and context to write something relatable.
-URL-to-Post: We use a custom summarization pipeline to extract the "hook" from any article or blog post and transform it into platform-specific formats (a technical post for X vs. a professional takeaway for LinkedIn).
So why now? We’ve just moved into beta after aggressive testing. We are looking for our first "real world" cohort of users.
The Ask: We would love for the HN community to break this.
-Does the conversational flow feel faster than a dashboard?
-Is the AI output "cringe" or actually usable for a technical audience?
-What’s missing from your current workflow that we should automate next?
We’re around all day to answer questions about the tech stack, the 12-month dev cycle, or our thoughts on the future of autonomous agents in marketing.