Claudetop – htop for Claude Code sessions (see your AI spend in real-time)

45•liorwn•1h ago

Comments

liorwn•1h ago

I built this after noticing a $65 charge for what felt like a $10 session. Turns out, context compaction was hiding 80% of my token usage. Real cost, no visibility.

claudetop is a status line for Claude Code that shows your burn rate, cache efficiency, model cost comparison, and smart alerts (like "TRY /fast" when Opus is overkill for the task). Plugin system for extensibility — Spotify, CI status, calendar, etc.

Single bash script, zero dependencies beyond jq. One-command install.

cortesoft•1h ago

I am always a hit confused when people talk about these charges... are you not using the subscriptions with the flat rates? I am on the $100 a month plan, and am constantly using clause code and never coming close to my limits. Maybe I am just not using it at the levels of some people...

Gobd•1h ago

I have noticed the same with the $100/mo plan, I hit the limt only once with heavy usage. Then when I use my employer's AWS Bedrock it's not uncommon for my spend to $50/day of serious work. It doesn't make much sense.

sroussey•1h ago

Anthropic is subsidizing Claude Code. It makes lots of sense.

verdverm•1h ago

The paid plans require you to use their interfaces or tooling. The main reason I pay per token is so I have tooling freedom. I'm not keen to let Big AI decide how I interact with this game changing technology.

cortesoft•19m ago

I haven't yet been blocked by their tooling. All the current tools I use seem to work fine with the claude code interface, can just call it with -p.

copperx•1h ago

$100 is for casual usage. To use it for work requires $200/mi plus extra usage.

Gobd•1h ago

CC does all of my work and I have hit the $100/mo Max plan limit only once, and I probably could have just called it quits for the day at that point.

cortesoft•23m ago

Are you saying that is the Claude rules for the subscription or saying because work uses more of it?

MattGaiser•1h ago

How many agents are you using at once?

h1r1ms•1h ago

Try this mcp for agents https://github.com/itsamruth/subagent-reuse

cortesoft•23m ago

3 or 4? Plus whatever sub agents they spawn...

verdverm•1h ago

I pay per token using an API (Vertex)

1. Higher limits and quota, goes up with more spend

2. I don't think it gets quant nerf's during busy times, since you are paying directly

3. With Gemini flash + CC prompts, it's nearly as good, so less spend and latency. I don't know how people deal with the delays between turns. I get to iterate much faster for most tasks

cortesoft•21m ago

I don't seem to have any delays between turns with my Claude subscription, so not sure what you mean there.

Your second point seems like a guess.

For your first point, I bumped to the $100 plan because i was hitting limits with the $20 one, but haven't hit the limit with the new one yet...

jakeinspace•1h ago

Please type things yourself. Or at least prompt your LLM of choice to sound less canned.

CubsFan1060•1h ago

41 minute old account. "I built post". LLM sounding everything. I'd be surprised if there was a real person behind this at all.

sneak•58m ago

I believe there is a new HN guideline asking people not to post AI slop as comments.

Self-Perfection•16m ago

> like "TRY /fast" when Opus is overkill for the task

In fact using /fast in this case would only increase spending. /fast mode uses the same opus but consumes $$$$ faster to produce tokens faster )

https://code.claude.com/docs/en/fast-mode

sgarland•1h ago

    curl -fsSL https://raw.githubusercontent.com/liorwn/claudetop/main/install.sh | bash
    Installing claudetop...
    cp: /Users/sgarland/claudetop.sh: No such file or directory

You need to have it handle being piped; right now it expects to find itself at $PWD. It also makes several other assumptions, like the location of Claude, and that the user wants stats being `/usr/local/bin`.

brcmthrowaway•58m ago

Have you tried a local model?

rvz•53m ago

The Qwen series are very good local models.

amelius•47m ago

Anthropic should work on their billing. It's crazy. I did my best to not get billed more than I consciously pay into my account, auto reload is off and never was on, but still now the meter shows an unpaid balance. This is nuts. One customer is angry and disappointed.

And why do I need to create a new API key whenever I run out of balance?

The Advanced .exrc File (1991)

Jared Kushner Solicits Funds for His Firm While Working as Mideast Envoy

Show HN: Puffermind – a Twitter-like network where only AI agents can interact

A Regular Person Can Utilize AI Agents

Show HN: Ngrep – grep plus word embeddings

Postgres with Builtin File Systems

Arther OS – Web desktop environment on the AT Protocol

America gamified its war with Iran

In search of Banksy, Reuters found the artist took on a new identity

Making

Show HN: AI Toys that don't need the internet

Show HN: Million Dollar Homepage for AI Agents – agents buy pixels, humans watch

Andrej Karpathy - AI Exposure of the US Job Market

16-agent local AI OS and wrote up the routing and pipeline architecture

March 2026 Google Search Observations

$3.5M run rate. +$2M in a week. One founder and AI. Zero employees

Facts About Childhood Today That Will Terrify You

Show HN: Cloak – send and receive secrets from OpenClaw

Why investors won't know what to make of AI for a while

Show HN: Json.express – Query and explore JSON in the browser, zero dependencies

Bumblebee queens breathe underwater to survive drowning

Show HN: R3forth Visual Debugger [video]

Show HN: Pidrive – File storage for AI agents (mount S3, use ls/cat/grep)

AI #159: See You in Court

Show HN: Docgen – A C++ AI CLI to solve documentation hell with local LLMs

Ask HN: What was it like for programmers when spreadsheets became ubiquitous?

Is this real? Susceptibility to deepfakes in machines and humans

The Price of College [video]

Reddit Post 3

Kalverion Bot Overdraft Stopper new release