I built a voice AI that speaks on your behalf — because I was frustrated from both sides of the same problem
I want to share why I built ReachMyAgent — because the reason is more interesting than the product itself.
I was frustrated from both sides
Side one — receiving inbound
During my job search I had back to back intro calls. Every single one started identically — "tell me about yourself", "what are you working on", "are you open to opportunities?" The same conversation, over and over.
Outside of that, Twitter DMs were piling up. Some I didn't have time for. Some I just wasn't in the right headspace to answer well. I'd send a rushed reply and immediately regret the impression I left. Some messages I never replied to at all.
Important connections were slipping away — not because I didn't care, but because I couldn't be everywhere at once.
Side two — trying to reach someone
When I want to learn about a person or a product, I have to piece it together from wherever I can find it. Portfolio pages. Landing pages. Twitter bios.
Landing pages especially give you a polished but often false image — everything looks great, nothing answers your actual specific questions. I just want to ask three questions and move on. But there's no standard way to do that. Sometimes there's a chatbot buried at the bottom of a page. Sometimes there's nothing. Every founder has implemented something different — if they've implemented anything at all.
There was no standard interface for "I want to quickly understand this person or product."
What I built
ReachMyAgent is that standard interface.
Every user gets one link — reachmyagent.com/username. Anyone who visits it can have a real voice conversation with an AI that speaks on their behalf. It knows their background, their work, their projects. It answers questions, handles meeting requests, and sends the owner a full summary after every call — with a pre-drafted reply ready to send in one tap.
From the owner's side — you never miss a connection. You wake up to summaries of every conversation that happened while you were away.
From the caller's side — there's finally one consistent, standard way to quickly understand a person or product. No hunting around pages. Just call and ask.
Where it is today
Free tier live — 30 minutes of calls
Connect GitHub, Codeforces, Kaggle, or upload a resume — agent context built automatically
Cal.com and Calendly integration for scheduling
Post-call email digest with transcript, summary, and one-tap draft reply
Try it
The best way to understand it is to experience it. Call my agent right now and ask it anything:
reachmyagent.com/[yourusername]
Would love feedback — especially from anyone who has felt either side of this problem.
Hello - I hope all is well in your busy world. I was immediately interested in your post, by reading the title of your posting. You have created a masterpiece. We all need this - we all want this. We have all experienced both sides of this as well. Thank you for all you are doing. Stay strong … enjoy the beauty of being YOU ☘️
Rooting for YOU
JUDITH ☘️
Such a clever solution to a very relatable problem
love how seamlessly it connects with existing tools like Codeforces to build context without extra work.
Congrats on the launch!
the "frustrated from both sides" framing is what makes this interesting. it's not just an outbound automation -- it's solving the coordination overhead that eats both sides of a conversation. I hit the same thing running async across timezones: the pre-call "tell me about yourself" loop is the most expensive part of the process and it produces zero new information. curious what the trust dynamic is like -- does the person on the other end know they're talking to the AI first? or is the reveal part of the design?
Hey, sorry for the very late reply, and thank you for appreciating my product, that I have created, and as per your question, the answer is yes. The other end knows that they are talking to an AI because we don't want to hide anything. The reveal is the part of the design.
yeah, owning the reveal makes it a feature, not a liability. tried the other way once and the backlash wasn't great.
True, most of the users feel cheated.
the "frustrated from both sides" framing is what makes this interesting. it's not just an outbound automation -- it's solving the coordination overhead that eats both sides of a conversation. I hit the same thing running async across timezones: the pre-call "tell me about yourself" loop is the most expensive part of the process and it produces zero new information. curious what the trust dynamic is like -- does the person on the other end know they're talking to the AI first? or is the reveal part of the design?
Congrats on the launch! Nice naming haha. Quick question: when you share updates about ReachMyAgent, do you post on just IH or adapt for Twitter/LinkedIn too? (asking because I built tool that handles the platform adaptation - running beta if useful)
Hey, yeah, I also post on Twitter every update about it.
Fellow voice AI builder here — really cool to see someone tackling the "be reachable without being always-on" problem. The two-sided framing (owner vs caller) is smart.
One thing I've learned building AnveVoice (voice AI for websites): the biggest unlock isn't just voice-as-interface, it's voice-as-action. When we added real DOM actions — the AI actually clicking buttons, filling forms, navigating pages — engagement jumped massively vs just answering questions.
Curious about your latency numbers. We spent months getting ours under 700ms end-to-end (STT → NLU → TTS) because anything above ~1 second feels unnatural in voice. What's the experience like on yours?
Also, how are you handling the summary quality? We found that shorter, structured summaries with specific action items convert way better than long transcripts. People want "they asked about X, seemed interested in Y, follow up on Z" not a wall of text.
The personal branding angle is interesting too. Most voice AI right now is focused on businesses (that's our lane — healthcare, e-commerce), but the individual use case for founders/creators has legs. Would love to hear how early conversations are going.
Hey, thanks for replying to my post and thanks for reading about it. Right now, the latency is very low. I have tried to use things as offline as possible, and I am less dependent on the APIs. I am using the services which have very fast inferences. That's what gives me the conversation speed, much better than Gemini Live API.
You built the interface for the gap between intention and follow-through. Most people miss connections not because they don't care, but because the friction of replying well feels heavier than the cost of silence. A voice agent that handles the first conversation and delivers a summary with a pre-drafted reply removes that friction without losing the human in the loop.
The two-sided framing is interesting but the payment asymmetry is the real test. Callers don't pay. Owners do. And right now an owner pays to have a voice AI wait for calls that won't come, because 'call my agent' isn't a behavior anyone has adopted yet. Calendly worked because senders already wanted frictionless booking. What's the wedge here? Is there an owner persona where inbound volume is already high enough that waiting for the standard to form isn't the problem?
You're right on the behavior adoption risk and I won't argue that.
The bet I'm making is that REACH-MY-AGENT doesn't need to be a universal behavior. It just needs to work in contexts where caller intent already exists. Someone who was going to DM a recruiter or reply to a founder's tweet — that's not a new behavior, that's a redirected one.
Where I think you're spot on: job seekers are probably the wrong wedge. They're mostly outbound, volume is low, the pain isn't there yet. The persona where this is immediately obvious is recruiters with open roles — 50+ applicants, DMs from sourcing campaigns, same 10 questions answered over and over. The inbound volume is already there, I'm just handling it differently.
Calendly parallel actually strengthens this for that persona: the pain exists, the callers are already motivated, the owner just needs a better drain.