❯ ojo --about
ojo = Observe · Warn · Offend(gated) # detect · alert · gated red-team
security that explains itself. an analyst in a container:
prevents in the PR · detects at runtime · investigates on its own.
deterministic bot + on-prem AI agent. your data never leaves.
# the critical decision is always deterministic; the AI only enriches.
❯ ojo problem
- ! nobody watches the logs · early-to-mid team with no security analyst
- ! data-leaking LLMs · "AI SOC" SaaS ships your logs (and PII) out
- ! alerts without context · someone investigates each one by hand
- ! rising incident cost · and nobody on call
❯ ojo how-it-works
- [01] prevent -> CI gate blocks the vuln before merge (auditable rule packs)
- [02] detect -> runtime: brute-force, enumeration, cross-tenant authz, geo, rate-spike
- [03] investigate -> read-only PII-free agent correlates and writes the report
# data-residency by default: on-prem model (Ollama) or hosted.
❯ ojo plans
freeunlimited CI gate$0
teamruntime + dashboard + alerts
business+ triage agent · integrations · SSO★ agent
enterpriseself-host · on-prem models · SLA
❯ ojo status
✓ 6 repos with the gate green ✓ 5 detectors (100% tested) ✓ on-prem agent
❯ ojo trusted-by
# teams & design partners running ojo
LOGOLOGOLOGOLOGOLOGO
❯ ojo contact --design-partner
closing 3-5 design partners (free, in exchange for feedback + a case study).
convergai@ojo:~$