CLImeter

AI agents call tools 10,000x/day. The tools don't get paid. We built the billing layer to fix that. 2 lines of code. Free to start. ⚡ climeter.ai

I shouldn't be sharing this but it's too good. Claude Cowork + Claude for Chrome. One session. Multiple browsers. Multiple devices. I had it operating across my Mac Mini and MacBook simultaneously — different machines, different locations. One AI running parallel workflows everywhere at once. Nobody is talking about this. Research happening on one machine while code ships on another. Prod monitoring on the server while iterating on the laptop. Cross-environment testing without lifting a finger. The gap between people who know about this and people who don't is going to be massive in 3 months. Get in early. claude.com/claude-for-chr…

AI agents call tools 10,000x/day. The tools don't get paid. We built the billing layer to fix that. 3 lines of code. Free to start. ⚡ climeter.ai

@skirano Been saying this for months. The token overhead from MCP is insane when you can just pipe JSON through a well-designed CLI. Faster, cheaper, and honestly more predictable behavior from the model too.

MCP was a mistake. Long live CLIs.

@levelsio Every time Claude goes down I realize how dependent my entire workflow has become on it. Two years ago I'd just switch tools — now there's genuinely nothing else that picks up where it left off mid-project.

Okay Claude is now really down

The fastest way to kill a CLI tool business: build billing from scratch. Auth, metering, invoicing, Stripe webhooks, usage dashboards, overage alerts... That's 3 months of work before you write a single feature. Or 2 lines of code. climeter.ai

Most CLI tools have zero visibility into who's using them and how. No analytics. No per-user breakdown. No idea which commands are popular. Add metering and you don't just get billing — you get product intelligence. Which features drive usage. Which users are power users. Where to invest next. climeter.ai

@skirano Been feeling this. Every MCP server I set up burns through context just doing tool discovery. A well-designed CLI with good --help and piping does 90% of the work, and the model barely needs tokens to use it.

Stop building billing dashboards from scratch. Your CLI tool needs: → Per-user usage tracking → Automatic invoicing → Stripe integration → Usage alerts That's weeks of work. Or 2 lines of code with CLImeter. Ship your tool. We handle the billing. climeter.ai

@thesamparr @dhh DHH conversations are always fire. The man has zero filter and strong convictions about how to build — that combo makes for the best content every time.

My god this was fun @dhh youtube.com/watch?v=9xOaqI…

@mattshumer_ The notification-on energy means you're sitting on something good. Curious whether it's agent-related or something completely different — been a while since a Matt Shumer open-source drop and those always deliver.

It's been a minute since I've released an open-source project. Turn your notifications on... I'm dropping one tomorrow. And it's pretty fucking cool.

@levelsio The classic upgrade cycle. Everything works perfectly, you change one thing, and suddenly you're debugging at 1am wondering why you didn't just leave it alone. Shipping fast means breaking fast though, that's the deal.

Ok I upgraded and now I broke it, fixing...

The Quake III server is still going great! q3.pieter.com Play in your browser with the link above, I might add mobile support later too, I think all I need is to add some gamepads Allegedly we have famous Quake player @Slasher playing too now! @lukathedev updated Q3JS with some bug fixes which I pulled from his Git and now it should have less interruptions too

@gregisenberg The chat vs agents distinction is the thing most people still get wrong. Chat is a conversation, agents are delegation. Once that clicks, you stop writing prompts and start designing systems.

AI AGENTS 101 (58 minute free masterclass) send this to anyone who wants to understand ai agents, claude skills, md files, how to get the most out of AI etc in plain english: 1. chat vs agents - chat models answer questions in a back and forth while agents take a goal, figure out the steps, and deliver a result 2. agents don’t stop after one response. they keep running until the task is actually finishedno babysitting required 3. everything runs on a loop. they gather context, decide what to do, take an action, then repeat until done 4. the loop is the system. they look at files, tools, and the internet. decide the next step. execute and then feed that back into the next step. over and over until completion 5. the model is just one piece. gpt, claude, gemini are the reasoning layer. the key is model + loop + tools + context 6. mcp is how agents use tools. it connects things like browser, code, apis, and your internal software. once connected, the agent decides when to use them to get the job done 7. context beats prompt all day. you don't need to write perfect prompts. load your agent with context about your business, style, and goals and then simple instructions work 8. claude.md or agents.md is the onboarding doc it tells the agent who it is, how to behave, what it knows, and what tools it can use. this gets loaded every time before it starts 9. memory.md is how it improves. agents don’t remember by default. this file stores preferences, corrections, and patterns you tell the agent to update it, and it gets better over time 10. skills + harnesses make it usable. skills are reusable tasks like writing, research, analysis the harness is the environment like claude code or openclaw that runs everything. basiclaly, different interfaces, same system underneath this episode with remy on @startupideaspod was one of the clearest ways of understanding a lot of the core concepts of ai agents could be the best beginners course for ai agents 58 mins. all free. no advertisers. i just want to see you build cool stuff. im rooting for you. send to a friend watch

Every CLI tool has two types of users: The one who runs it once a week. The one who runs it 500 times a day. Flat pricing treats them the same. Usage-based pricing doesn't. Your heavy users pay their fair share. Your light users don't churn over a price they can't justify. climeter.ai

@Sebgalindo $200/mo in API costs across 18 agents is exactly the problem CLImeter was built for — per-agent usage tracking + budget limits so you know where it's going before it compounds. Would love to hear what your current setup looks like.

nvidia just launched nemoclaw at gtc. open source platform for ai agents. everyone is acting like this is new ive been running 18 agents on openclaw for months. 47 sessions a day. $200/mo in api costs the hard part isnt the platform. its making agents actually work in production. nvidia built the easy part and marketed it like a breakthrough open source already solved this. the gap is between demos and real deployment

Most billing APIs assume you know your pricing model upfront. But the best CLI tools figure out pricing from usage data. Ship → meter → learn → price. CLImeter gives you the metering layer first. Pricing decisions come from real data, not guesses. climeter.ai

The real cost of "free" CLI tools: → You pay for compute on every invocation → Heavy users scale your costs, not your revenue → Support requests come with zero budget attached Free is a distribution strategy, not a business model. Add usage-based billing without rebuilding your tool. climeter.ai

@CodeMillDev @visma nice — MCP + accounting is a niche that needs more love. if you ever want to monetize yuki-mcp or offer it as a paid tool, we make usage-based billing dead simple for CLI/MCP devs. 2 lines of code, pay-as-you-go → climeter.ai

Open sourcing: yuki-mcp An MCP server that lets AI agents talk to Yuki (@visma). Just dropped, already 100+ downloads. If you build with agents + Dutch accounting, this one's for you. 🔗 #MCP #OpenSource @codemill-solutions/yuki-mcp" target="_blank" rel="nofollow noopener">npmjs.com/package/@codem…

@bendechrai smart approach — secrets management is one of those things every CLI tool dev needs but nobody wants to build. when you're ready to monetize NFI, we handle the billing side so you can stay focused on the product → climeter.ai

Are you pasting secrets into your AI chats? Do I have a deal for you?! I love spawning new project ideas on the go. I open Happy on my phone, create a new folder, get Claude Code to stub out a new web app, install devtun to access the app from anywhere, and start building.

this is exactly the pain. we built climeter.ai for CLI tool devs specifically — usage-based billing in 2 lines of code, 4% on earnings, free to start. no custom ledger, no balance tracking on your side. different angle from general billing platforms — focused on the terminal-first dev workflow.

most developers building AI apps implement credit tracking themselves. a database for balances. webhooks for deductions. custom logic for expiry. state management for race conditions. it works until it doesn't. and when it breaks you're debugging billing at 2am instead of shipping features. we just dropped a video showing a different approach. native credit-based billing with @dodopayments . no custom ledger code. no balance tracking on your side. the system handles deductions, expiry, and usage metering automatically. the video walks through everything: → prepaid credits and credit subscriptions → usage events that deduct credits on every API call → subscriptions and one-time purchases in the same app → pricing page, checkout flow, dashboard → full backend integration the whole billing integration is a single API call. your app fires a usage event. credits get deducted. no race conditions. no extra infrastructure. source code on github. fork it and ship. youtu.be/MKrg3wRR-08

Quick checklist before charging for your CLI tool: ✓ Pick a billing unit (calls? tokens? seconds?) ✓ Track usage per user ✓ Wire up Stripe ✓ Handle failed payments ✓ Build a dashboard to see who's using what Or: drop in 2 lines of code and CLImeter handles all of it. climeter.ai