Partners
Working with NOPE
Every partnership is shaped around the partner: what they run, who their users are, what their safety team is trying to do. Here is what that has looked like so far, and what holds in every case.
What it has looked like
Self-hosting, with us on call
Platforms run Edge or Ocular OSS inside their own stack under open licenses. When they want more than weights, we help: operating thresholds, calibration quirks, known failure modes. The licenses don’t require the conversation, and we don’t ask who you are before helping.
On-prem delivery
Packaged classifier builds for a customer’s own hardware, delivered through our portal with licensing and support sized to fit.
Design partnership
A small number of partners whose real conversation domains tune our detection: their personas, their user populations, their failure modes. What flows back to us is calibration and taxonomy under a written agreement. Conversation data never leaves the partner.
Research and policy
Researchers and policy teams use the incident record, the regulation tracker, and the open benchmarks for their own work. That use is free by design. If the evidence would serve your study or your rulemaking, talk to us.
What holds in every case
Your users’ conversations stay yours
Self-hosted deployments send us nothing. Cloud calls are processed without being stored. Nothing a partner shares trains a public model without a written agreement.
Our numbers can’t be bought
Results on the public test suites and open benchmarks are unaffected by commercial relationships, and no one can pay to re-run or improve a score. If that independence is why you’d trust us, it’s also why it’s not negotiable.
The open work stays open
Released weights stay released. The trackers, the crisis-resource directory, and the benchmarks stay free for customers and for platforms that will never pay us.
Start the conversation
Tell us what you're building and where safety fits. A short email is plenty. The same door works for researchers, policy teams, and funders.