Truly Typed – A writing app for the AI era
Hey HN, this is deepan from trulytyped (https://trulytyped.com). I am building a document writing app which makes it extremely easy to figure out how a document was created. Now that any text can be AI generated, how do you tell if something was actually generated or composed. It is impossible to detect AI after a piece of text has been generated. No amount of watermarking, linguistic checks or vibe checks work consistently. The AI detectors that schools and journals use are easy to bypass. Why do we need to solve this problem - First of all, this is not an anti-AI stance. I have…
What it does
In the maker’s words, at launch
Hey HN, this is deepan from trulytyped (https://trulytyped.com). I am building a document writing app which makes it extremely easy to figure out how a document was created. Now that any text can be AI generated, how do you tell if something was actually generated or composed. It is impossible to detect AI after a piece of text has been generated. No amount of watermarking, linguistic checks or vibe checks work consistently. The AI detectors that schools and journals use are easy to bypass. Why do we need to solve this problem - First of all, this is not an anti-AI stance. I have trained machine learning models in healthcare, cybersecurity and privacy space in last 10 years and open sourced some of the stuff (https://github.com/deepanwadhwa). I think, at a very fundamental level, we need to figure out how to differentiate human experience from AI so that we could have continued access to humans. Access through in-person, writing, audio or video. Access through text has already been fractured, our radius of trustworthy written material on the internet has reduced. And I think a similar trajectory will follow for audio, video and in person. So how does truly typed solve this - Each document composed in truly typed carries information such as how much of the content was actually typed, how much was pasted, how many sources were used, how many authors contributed. Each document also carries flags such as verified human, bot detected, unverified. Unverified is for cases where humans are transcribing (basically copying by typing) text looking at another tab or book. The core thesis of the product is that purely human generated thoughts will matter more and more. Since I come from a privacy and security background - Each and every profile and post is private by default. You decide if you want to make it public or not. We also have a pretty tight bot and automation defense - if you are an automation expert - try pointing a script at the app and see if you can get a verified human flag on an article (please reach out to [email protected], if you actually break it). We don't want to use your data for any llm training and we don't want to sell your data to any brokers. Our primary market for now is Academic Journals, News media outlets and Colleges and general folks who just wanna write and share with their audience. We are sort of unique in this offering and competition exists at different layers. Google docs and Microsoft word are both writing apps, but they don't really tell you if the document was actually created by a human or not. We pointed a bunch of our testing scripts at google docs and full articles generated with real keystrokes without google detecting a hair. There are a ton of AI detectors that we have come across which show how much of the text is human or AI but they all are easily bypassable. I am passionate about this problem and would like to hear your feedback, criticism, interest. Cheers, PS: I typed the whole thing myself. On truly typed, you won't have to say this.
Does the same job
all alternatives →
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, May 2026
the whole month →
Parallel agents, diff reviewer, and multi-model comparisons
Dev tools · May 2026 · kilo.ai
- NW
Hey HN, Henry here from Cactus. We open-sourced Needle, a 26M parameter function-calling (tool use) model. It runs at 6000 tok/s prefill and 1200 tok/s decode on consumer devices. We were always frustrated by the little effort made towards building agentic models that run on budget phones, so we conducted investigations that led to an observation: agentic experiences are built upon tool calling, and massive models are overkill for it. Tool calling is fundamentally retrieval-and-assembly (match query to tool name, extract argument values, emit JSON), not reasoning. Cross-attention…
Life & fun · May 2026 · github.com
- FM
Dev tools · May 2026 · github.com