HN Showcase
Back to browse
GitHub Repository

High-throughput inference for your agents — run many of them in parallel over your own private data, so you pay for your context once, not on every turn.

36 starsPython

Claude Code's $200 plan is a 17× subsidy on the raw API

by Hiteshjain118·May 27, 2026·9 points·23 comments
Visit ProjectView on HN

AI Analysis

●●SolidBig BrainDark Horse

150× token inflation finding — you pay to re-read, not generate.

Strengths
  • •150× token inflation finding is concrete, measurable, and genuinely surprising for users.
  • •Local-only analysis reading ~/.claude logs respects privacy, nothing leaves your machine.
  • •Breaks down actual cost driver: 84% of bill is input tokens, not generation output.
Weaknesses
  • •Repo is sprawling with multiple packages — lacks focus on the token xray insight.
  • •Only works for Claude Code users, limiting audience to one tool's userbase.
Category
AI/ML
Target Audience

Claude Code users, AI developers tracking LLM costs

Similar To

LLM cost trackers · Token usage analyzers · Claude dashboard tools

Similar Projects

Developer Tools●●Solid

Ledger – Claude Code Token Spend Analyzer

Local SQLite ledger captures runaway Claude Code sessions before the bill shocks you.

Solve My ProblemCozy
tsv650
301mo ago
Security●●Solid

Agent-pd – A zero-token audit log to catch rogue Claude Code subagents

Catches denied calls that never reach transcripts—six detectors at zero token cost.

Big BrainNiche Gem
softie123
629d ago