Inference-native Tokenmaxxing Agent Harness for Loop Engineering
EvE is an open-source framework for co-evolving ensembles of coding agents. Wrap Codex into an evolutionary loop that co-evolves solvers and agent guidance.
Leave things better than you found them! This claude skill seamlessly integrates iterative feedback across your claude stack and keeps an audit log/paper trail.
Privacy first toolkit for analyzing your ChatGPT and Claude exports. Model tracking, prompt effectiveness analysis, cost efficiency metrics (POD/DLOD/POE), topic classification, and industry benchmarks. 49 tests. pip installable. Everything runs locally.
Tokenmaxx or log off. See what your friends are building in the Claude Code statusline, and who burns the most tokens today. Prompts never leave your machine.
Convert PDFs, Office docs, images & web links to clean, token-efficient Markdown — 100% locally. Powered by Microsoft MarkItDown.
Display Claude.ai usage limits on a $30 desktop WiFi clock (GeekMagic SmallTV ESP32)
Is AI a bubble? An essay on why AI is both the most transformative technology of our lifetime AND a historic bubble — tokenmaxxing, the $500M Claude accident, OpenAI's losses, and the circular hyperscaler money.
transparent TOON re-encoding for AI coding agent tool-call context
A header only library for driving a voodoo 3 from userspace. Created by fable, verified on hardware by me
Repo-intelligence distributes and integrates three complementary context-economy tools across multiple projects.
Like a utility meter for your house, but for AI API consumption
Parasail is an AI Supercloud — a pay-per-token GPU inference platform aimed at AI startups and developers. Parasail orchestrates rented GPU capacity across 40+ data centers in 15+ countries to serve open-weight LLMs, vision/multimodal models, embedding models, and TTS/STT models on a serverless, dedicated, or batch basis.
Get more out of lower-cost settings/smaller LLMs by utilising your terminal for local agent-like workflows - via this lightweight safety-focused AI harness that uses bounded scripts, human approval gates, compact reports, & deterministic routines - all in one stack. For tokenmaxxing, and for reducing shell-risk, drift, & hallucinated local actions
Chrome extension that improves Claude.ai's usage limits page showing your pace in top of your usage. Tokenmaxx your sub.
Brock Denton's public profile and projects, including Tokenmaxxing.
Local-first context compression for Codex CLI and Claude Code CLI on macOS. Shrink old context, protect exact details, and recover originals safely.
Autonomous engineering memory for coding agents. Captures diagnostics, git diffs, terminal evidence, and project context to provide persistent memory and continuity across development sessions using HydraDB.
Tokenmaxxing: a reproducible benchmark of 5 levers (fine-tuning, RAG, prompt-stuffing, prompt optimization, model tier) for LLM agents on Azure OpenAI, measured by quality-per-token, total cost (TCO) and latency. Spoiler: base + RAG wins.
Add a description, image, and links to the tokenmaxxing topic page so that developers can more easily learn about it.
To associate your repository with the tokenmaxxing topic, visit your repo's landing page and select "manage topics."