playwright MCP (73% less tokens). Quick setup: npx playwright-slim --setup
memory MCP (44% less tokens). Quick setup: npx memory-slim --setup
30 tools vs 1: an agent-payment MCP that measures 230 tokens of schema floor instead of 3,705 (16x smaller; ~48x effective across a multi-hop payment flow). Signed, sanctions-screened, trust-gated payments across x402, Stripe, AP2, MPP, Mastercard Agent P
Real-time cost awareness for MCP agent workflows
filesystem MCP (54% less tokens). Quick setup: npx filesystem-slim --setup
notion MCP (52% less tokens). Quick setup: npx notion-slim --setup
git MCP (58% less tokens). Quick setup: npx git-slim --setup
An MCP server to analyze AI chat sessions and calculate token wastage using git repository analysis.
OpenCode plugin for web search and git clone — gives agents access to the outside world
Tokenizer-aware chat-history fitter. Trims, windows, or summarizes old messages so your conversation fits the model's context budget — with a full audit log of every kept, dropped, and summarized message.
Fast heuristic token estimator for LLM prompts. Zero dependencies, no tokenizer downloads. Within ~15% of real tokenizer counts for prose.
Token-efficient AI coding agent configuration. Generate optimized CLAUDE.md, .cursor/rules, AGENTS.md and more from a single source of truth.
WebAssembly bindings for bpe-openai tokenizer
Circuit breaker for AI agents — pick budget-guard or loop-killer mode and stop runaway token spend or stuck agents in one wrapper. Adapters for LangChain, OpenAI Agents SDK, and the Claude Agent SDK.
Markdown document analyzer for AI agents - extract metadata, headings, links, tables, tokens, and key points
Lightweight, production-grade token usage tracker and rate limiter for AI API calls.
A CLI tool to recursively compile files into a single markdown file.
A Pi extension that indexes your local TypeScript/TSX codebase into Chroma Cloud, enabling semantic and symbol-based code retrieval directly within Pi's agent loop.
Measures the token weight of instructions, MCP servers and MCP tools for AI agents
CLI for running Quadtrix.cpp chat and local training.
Token-aware prompt minifier with explainable diff. Shrinks LLM prompts and tells you exactly what changed.
The `tokenizer` function uses the `js-tiktoken` library to encode the input string into tokens using the GPT-2 encoding scheme. It then decodes the tokens back into strings, maps the tokens to their positions in the input string using the `mapTokensToChun
Cloc for counting tokens
Estimate the number of tokens an OpenAI chat completion request will use