Hey Product Hunt! I'm Leo, the maker of Ahy Governance. The problem: Last month I deployed 5 AI agents to handle contract review. Within 3 days: - Agent A said "deadline is June 30" - Agent B said "deadline is July 15" - Nobody noticed until a client flagged it - The token bill hit $200 with zero visibility into which agent cost what I looked for tools to fix this. LangSmith and LangFuse are great — but they trace individual LLM calls. They don't understand multi-agent orchestration. They can't detect when two agents produce contradictory outputs. They can't tell you how much each agent costs. So I built Ahy Governance. 7 modules, 312 tests, MIT licensed. What makes it different: 1. Conflict Detector — 5 types of inter-agent conflicts (fact, format, dependency, scope, confidence). Currently unique in the market. 2. Cost Tracker — per-agent attribution with 22 model pricing tables and budget circuit breakers 3. Audit Logger — SHA-256 hash chain, tamper-proof. Export SOC2 or ISO27001 compliance reports with one click. 4. Health Monitor — heartbeat, P50/P95/P99 latency, pipeline DAG tracking 5. RBAC + API Keys — 3-tier roles, multi-tenant isolation 6. Prompt Guard — 13 injection patterns + PII redaction 7. Memory Sharing — namespaced key-value store shared across agents Pricing: Community (free, 1 agent) → Pro ($99/mo, 10 agents) → Team ($299/mo, 50 agents) → Enterprise ($999/mo, unlimited). Tech stack: Python >=3.10, FastAPI backend, single-file HTML dashboard. Zero npm, zero build step. `pip install ahy-governance[web] && ahy-dashboard` and you're running. I'd love feedback on: - What's missing before you'd use this in production? - Which module should I tackle next? - Would you pay for this? If not, what would make it worth paying for? Thanks for checking it out!

GitHub

Detect agent conflicts. Track costs. Audit everything.

Detect agent conflicts. Track costs. Audit everything.