wanory.com is live.
Wanory Labs' official site is live: meet Inwit-1 in open beta, explore developmental AI research, and contact the lab.
By Jiilan Nashrulloh Tanjung, Founder, Wanory Labs
What this site is
Wanory Labs finally has a home on the internet. Until now the lab lived in chat windows, API consoles, and email threads. This site gathers everything in one place: the research direction, the live product, the specs with honest limits, the invest case, and a way to reach the person building it. Nothing here is a teaser. Every page states what exists today, what is still in development, and what happens next. If you are trying Inwit-1, start with the product pages. If you are evaluating the research, read how Inwit-1 grows first. If you are considering the first check, the invest page lays out the snapshot, the thesis, and the use of funds. The site will grow as the lab grows, and this article stays as the record of day one.
How Inwit-1 grows
Inwit-1 does not come from a training run. It starts from a digital genome, a compact encoding that describes how a mind can develop rather than a finished set of weights. From that genome, cells divide and circuits wire themselves into a young mind, the way biology builds brains: structure first, then experience. Every conversation shapes it further. When Inwit-1 serves you, the session changes the mind that serves the next one, so the system keeps developing instead of freezing on the day training ended. This is the whole wager of the lab. Prediction scaled further than anyone expected and still hits the same ceiling, because a frozen model cannot keep learning from the world it serves. Development keeps the structure plastic. The product pages describe the consoles, and the research direction runs through all of it.
Two products: one live, one building
Two products live on this site, at very different stages. Inwit-1 is live in open beta and free during beta. Open /chat/try in your browser: no signup, no request, no card. Chat works immediately, and the API and Brain consoles open through access requests while the beta runs. There is no SLA yet, and the specs publish limits next to abilities, because safety before scale is the lab's first principle. WanoryCode is the second product and it is still in development. It is an agent harness for software work: describe the change in plain words in the terminal where you already work, and the agent reads the repo, plans the change, edits files, and runs commands. Nothing lands without you. Every change arrives as a diff you accept or send back. Early access runs over email while we build.
Predicted minds versus grown minds
A comparison page on this site lays out the difference the lab is betting on. Transformer models learn during training runs, then freeze. When they are wrong, the fix is another run: collect data, retrain, redeploy. Inwit-1 learns by developing. When it is wrong, the experience itself reshapes the structure, the way a person adjusts after a mistake rather than being rebuilt from scratch. To improve a transformer you rerun training. To improve Inwit you keep using it. The page puts this side by side across four aspects: how each learns, when learning stops, what happens when wrong, and what improvement costs. It also answers the three questions readers ask most. Read it if the research claim sounds too big. It is written to be argued with.
The lab behind it
Wanory Labs is an independent AI research lab in Indonesia, founder-led and solo by design. Jiilan Nashrulloh Tanjung founded the lab and leads research on Inwit-1. Small on purpose means fewer meetings and more development: the lab ships Inwit-1 in open beta while staying lean, and every message is answered by the builder. Three principles govern the work. Safety before scale: limits are published next to abilities and releases stay careful. Useful over impressive: the lab would rather ship something people rely on than something that demos well. Open access in beta: Inwit-1 is free to try with no signup while the beta runs. The lab is pre-seed with no investors so far, and the full case for the first check is public on the invest page: snapshot, thesis, use of funds, and terms discussed directly with the operator.
Start here, then talk to us
Different visitors need different paths, so here is the short map. Trying the product: open /chat/try, then read the Inwit-1 specs to know the limits. Evaluating the research: read how Inwit-1 grows, then the comparison page. Considering investing: the invest page has the snapshot, the thesis, and the use of funds, then write to hello@wanory.com. Joining the lab: the leadership page describes how the lab works and who it is hiring across research and engineering. Covering the lab: send a press inquiry to the same address with the subject Press inquiry, and brand assets come on request. One address for all of it, answered by the person building it. This site is the front door. The conversation happens over email.
What happens next
The beta decides what gets built next. Access widens from chat toward the API and Brain consoles as the beta grows, and pricing after beta will be decided later and stated plainly. WanoryCode moves from development toward early access, shaped by the requests that come in over email. The invest page stays current as the round takes shape. Everything on this site is version day one: pages will be updated, specs will change as limits move, and this article stays frozen as the launch record. Bookmark the newsroom. Product ships and company milestones get published there, nothing else.