Sidegrade — AI skill for Claude Code
Same intelligence, less money — reads your local Claude Code / Codex / Hermes / OpenCode usage and shows which model gives the same intelligence for less.
How to install Sidegrade
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open iotexproject/sidegrade and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Sidegrade does
Same intelligence, less money — reads your local Claude Code / Codex / Hermes / OpenCode usage and shows which model gives the same intelligence for less. 100% local, npx, no account.
Alternatives in AI
- Codeburn — Free, local tool to track AI coding token usage and cost across 37 tools and agents (Claude Code, Cursor, Code 9.7k ★
- Surface — Show ranked attack surface for a target from its recon manifest + hunt memory 3.8k ★
- Token Monitor — Local-first desktop widget for tracking token usage, costs, and limits across 32+ AI coding tools—including Cl 1.8k ★
README
sidegrade
**Same intelligence, less money.** `sidegrade` reads your local coding-agent usage — **Claude Code, Codex, Hermes, and OpenCode** — and tells you which model gives you the *same intelligence* for a lower price, and which cheaper tier is worth a small step down.
npx sidegrade
No install, no account, no config. It reads your existing Claude Code, Codex, Hermes, and OpenCode session logs on this machine and prints a report — whichever of the four you use.
Or just ask your agent
Already inside Claude Code, Codex, Cursor, or any coding agent? Paste this — it runs sidegrade and tells you what to do, no terminal needed:
Run `npx -y sidegrade --json` and tell me the intelligence (AA) I'm running and
what it costs per day, the cheapest model at the same intelligence, and a cheaper
step-down — with $/day and savings for each.
What you'll see
Run it and you get a report built from *your* actual usage. Example:
sidegrade · same intelligence, less money
100% local · no account · no telemetry · your prompts & code never leave this machine
You're using AI at intelligence AA 53, costing about $151/day at list price.
Based on 3.3B tokens over the last 30 days · 94% cache-hit · Claude Code (26d), Codex (21d)
Intelligence-per-dollar (if you ran all of it on one model)
model AA vs you $/day via
────────────────────────────────────────────────────────────────
Claude Fable 5.1 57 +4 $201 QuickSilver Pro same+ ◆
GPT-6 Astra 55 +2 $326 NEAR AI same+
Claude Opus 5 54 +1 $130 QuickSilver Pro same+ ◆
Muse Spark 1.3 53 -0 $29.21 QuickSilver Pro same+ ◆
Grok 4.6 51 -2 $81.50 NEAR AI ~-2
Kimi K3 50 -3 $83.05 QuickSilver Pro ~-3
GLM 5.3 49 -4 $36.93 Z.ai ~-4
Gemini 3.8 Flash 47 -6 $20.76
Related Skills
Tokimeter
Local-first usage & cost meter for AI coding agents: Claude Code, Codex, Cursor, Grok Build, Hermes, opencode,
Core
Agentic AI monitoring and observability — OTEL traces and local log files from Copilot, Claude Code, Codex, an
Agentlens
Agentic AI monitoring and observability — OTEL traces and local log files from Copilot, Claude Code, Codex, an
AI Observibility
Local-first behavioral observability for AI coding agents. Reads your Claude Code transcripts and shows whethe
Usewarden
Free, local guardrail for AI coding agents — Claude Code, Cursor, Gemini CLI and more. Blocks out-of-scope wri
Skill Cost
Multi-harness token-usage cost reporter for AI agent sessions (pi, Claude Code, Codex, opencode). Reads JSONL
Related Agents
Cost Analyst
Analyzes AI spend and budget via PromptWise - cost breakdown, forecast, and routing savings. Use when asked wh
CLI Engineer
Ingénieur du CLI Hookstack (packages/cli/, package npm hookstack-cli). Utiliser pour toute évolution du CLI in
Ensemble Curator
High-volume triage agent for external-AI-agent PRs (Codex, Jules, Hermes, Droid, Aider, etc.). Analyzes, conso