a lane is a rules.json or .head pack — write your own (RULEPACK.md)

Moderation policy layer (score-consumer)

free · certified

mod_policy@0.1.0 · 4 routes

The decision layer over YOUR scorer: post category scores {nudity, violence, drugs, self_harm, scam} from any upstream model — SightEngine, SigLIP, openmod, your own ONNX — and get a calibrated publish/review/remove/unchecked verdict. Threshold policy is a rules file you own; absent inputs escalate instead of silently passing.

Try it live — pinned to expert:"mod_policy"

Free demo endpoint, no signup. Same call your agent would make.

Taxonomy (4 routes)

publish, review, remove, unchecked

Honest limits

Demo thresholds are hand-authored — production deployments tune the rule file to their own scorer's score distribution (that's the €149 custom-lane job). This is the honest image-moderation path: your scorer does pixels, mtlm owns the typed verdict.

Install

python3 tools/lane_install.py mod_policy@0.1.0 --apply --appliance .
python3 tools/lane_repro.py mod_policy-0.1.0.lanepack.tar.gz --url http://localhost:8402   # re-run the published eval yourself

sha256: 6c16cb900b03a880bdc14156… · Apache-2.0 · download lanepack

NEED YOUR OWN ROUTES?

A private lane on your taxonomy — trained, calibrated, exclusive. €149 / domain, 3 business days.

Get a lane like this pricing →

generated by tools/render_lanes.py · source: index.json