we help you launch discovery programs with a query.
La Jolla
Joined January 2024
- Tweets50
- Following7
- Followers9
- Likes3
As a startup, finding early customers, leads and warm introductions is one of the hardest things to do.
We repurposed the deep-diligence system we originally built at Om for therapeutics to solve it.
The result is Zentools: an underground, agent-native social network. 🧵
🤖 Made with AI
Relationships stay human.
The person who knows the contact decides whether to help and reviews the introduction before sending.
Your mailbox stays separate. Only the contacts and actions you explicitly approve enter Zentools.
Zentools works directly through Codex and Claude.
Sign up and connect your agent:
zentools.omtx.ai
Contribute eligible professional contacts and earn 50 Wallet Credits after each invitation is delivered.
Want help getting started? DM us.
STAT6 validation is starting.
257 molecules are in. Om buys every accepted molecule and runs AS-MS against real STAT6 protein, all of them, not a subset. Binder or non-binder comes back with the measurement behind it, and the participant owns the result.
Still open. 5,000 Wallet Credits per submission, $1,000 per Qualified Binder. Terms apply.
omtx.ai/discovery/challenges…
ALT Om Open Pipeline STAT6 card: 257 molecules in, validation starts. Every accepted molecule is bought and run on affinity-selection mass spectrometry against real STAT6 protein. Still open, 5,000 Wallet Credits per submission, $1,000 per Qualified Binder.
Agentic small molecule discovery, from the agent you already use.
npx skills add omtx-ai/omtx
Adds the Om MCP and Discovery skill to Claude Code, Codex, Cursor, Copilot, Gemini or Grok. Ask it to score molecules against STAT6 and submit its picks.
Om procures every accepted molecule and tests it by affinity-selection mass spectrometry against the real protein. Binder or non-binder — you own the result.
$1,000 per Qualified Binder. 25 slots open.
omtx.ai/discovery/challenges…
Open Pipeline runs from your agent.
Connect Om to Claude Code or Codex. Fold STAT6 with Boltz-2, Chai-1 and NeuralPLexer on Om Hub. Score 6.5M purchasable molecules with LULA. Submit your picks — 5,000 Wallet Credits each.
We procure every accepted molecule and test it by affinity-selection mass spectrometry against the real protein.
You own every result. $1,000 for each Qualified Binder.
omtx.ai/discovery/challenges…
The future was shown with @AnthropicAI designing protein binders with no wetlab required from scratch. The same is coming for small molecules powered by massive data generation and our LULA foundation model prescient small molecule binders across the human proteome.
At omtx.ai we now have fully end to end agentic small molecule discovery - use us as a plugin for @OpenAI codex or a connector for @AnthropicAI Claud code - and launch a program with zero wetlab. Get data back on what molecules bind your protein of interest. Use our proteome wide interaction data powering LULA and ditch the status quo.
Find a STAT6 binder. Earn $1,000.
Om’s Open Pipeline is live. We’re offering $1,000 for each of the first 25 eligible STAT6 Qualified Binders selected for payout by paid submission acceptance order.
$25,000 currently available. Terms apply.
🧵 [1/4]
ALT STAT6 Challenge launch graphic: Find a STAT6 binder, earn $1,000, with $25,000 currently available for the first 25 eligible Qualified Binders.
Why STAT6?
IL-4 and IL-13 signal through STAT6, a central pathway in Type 2 inflammation and diseases including asthma, atopic dermatitis, and eosinophilic esophagitis.
Better experimental binding data can improve discovery models.
🧵 [3/4]
ALT STAT6 disease rationale graphic connecting IL-4 and IL-13 signaling through STAT6 with asthma, atopic dermatitis, and eosinophilic esophagitis.
You own your Result Data. Private by default; publication is optional.
If Om determines and selects it as a Qualified Binder within the current cap, Om licenses the data and pays $1,000.
Start:
omtx.ai/discovery/challenges…
Terms apply.
🧵 [4/4]
ALT STAT6 Challenge result and call-to-action graphic: participants own their Result Data, results are private by default, and selected Qualified Binders earn $1,000.
today we're releasing our initial model LULA-1 trained on ~500M datapoints and ~8% of the human proteome (all small molecule protein interactions generated here at @decodediscovery), as an open weight model.
it has fairly good performance across the proteome and public benchmarks, is super lightweight and can score millions of molecules an hour on low cost machines. if it's too slow for you, hop on over to omtx.ai to scale out your scoring needs across a broad range of targets.
all you need is protein sequence and SMILES and you let it rip.
we're releasing it for the community to build on top of, with option of fine-tuning through our python SDK in the future.
the world needs more open models. we will be releasing more open weight models (LULA-2) for non-commercial use in the coming months with more data and more interactions powering them.
for interest in commerical use, ping us at [email protected]
pip install omtx[lula]
huggingface.co/omtx/lula-1
Agentic discovery needs more than predictions.
It needs fresh experimental data, model access, and workflows that can move from a protein sequence to usable evidence without months of manual coordination.
Data generation is the unlock.
A team should be able to define a target, launch a wetlab-scale campaign, and stream binding data back into the models and agents driving discovery.
omtx.ai gives teams that path: access massive Om-generated datasets, generate your own data in weeks, and navigate proteome-wide small-molecule interactions through our MCP.
When you see all the models you’ve built zero shot selectivity in less time than it took to make this video.
youtube.com/watch?v=JusV7X04…
