Licentric
Control plane for AI agents: identity, budgets, cost
Start new thread

Licentric - Control plane for AI agents: identity, budgets, cost

by•
Licentric is the control plane for AI agents. Give every AI Agent an identity, hard spending budgets, governance policies, evals, traces, and per-model cost attribution. One REST API, Python and TypeScript SDKs. What makes it different: it is built on a real licensing engine. The Stripe-to-signed-key pipeline, offline Ed25519 validation, and device fingerprinting that protect paid software now authenticate, meter, and bill autonomous Agents. From license keys to agent tokens, in one platform.

Add a comment

Replies

Best
Maker
📌
Hey Product Hunt and fellow builders 👋 I'm Able, founder of Licentric. Which AI Agent is running right now? What is it allowed to spend? What did it actually do? If you run AI agents in production, you probably can't answer all three with confidence. I couldn't, and that's why I built Licentric. Licentric started life as something unglamorous: a software-licensing platform - license keys, device activation, offline validation, Stripe-driven provisioning. Solid, boring infrastructure. Then I started shipping AI agents, and that boring infrastructure turned out to be the missing piece. Per-seat licensing is meaningless for an agent that runs 10,000 times overnight, and "just give the API key to the model" is how teams wake up to a runaway bill with no way to prove what happened. Here's the insight that became the product: an autonomous agent is just a non-human identity — and identity, entitlements, hard limits, metering, and billing are exactly what a licensing engine already solves. So I didn't bolt on another dashboard. I built the agent control plane directly on top of the engine that already secures paid software: 🪪 Identity + lifecycle for every agent (active / suspended / revoked) 💸 Hard token & spend budgets — per agent, team, and workflow 📋 Governance + an immutable audit trail (built with the EU AI Act in mind) 🧪 Evals to score agent quality before you ship 🔍 Traces to see exactly what each agent did 📊 Per-model cost attribution across providers One REST API. Python + TypeScript SDKs — pip install licentric, npm i /sdk. The same Ed25519-signed, offline-first crypto that protects licensed software, now protecting your Agents. The regular free tier is generous and forever: 100 active licenses, 5 agents, 1,000 API requests/mo, 1 product — no card, no trial clock. 🎁 Founding Hunters — the first 100 only. Comment "PH" + your signup email and I'll personally lock you into 1,000 licenses + 50 agents, free FOR LIFE. Not a year. Life. When 100 are claimed, it's gone for good and won't come back. Miss the window, sign up on standard terms. Here's the uncomfortable part: right now your agents are spending money you can't see, can't cap, and can't explain after the fact. That doesn't get safer the longer you wait — and with the EU AI Act landing Aug 2026, "we'll add governance later" stops being a plan. I genuinely want to hear how you handle agent identity, budgets, and cost today — and where Licentric falls short. I'm in the comments all day. Say "PH" and I'll get you set up before the window closes. Thanks for taking a look 🙏 - Able Varghese