whereismyai
Token Beggar
Token Beggar is a coding agent plugin for Claude Code that routes each task to a cheaper model and effort level.
Token Beggar is a plugin for Claude Code that checks whether a cheaper model can handle a coding task just as well before the main model does the work. Small tasks stay in the chat, while larger tasks like fixing an unfamiliar repository go to a cheaper worker while the chat checks the diff and runs tests. It runs locally on your machine in your Claude Code session, uses your existing sign-in, and writes only to its own folder. It supports Python 3.10 or newer and requires Claude Code. Measured results show it reduces cost on hard tasks while resolving a similar share of issues compared to working alone.
Token Beggar at a glance
Key features
- Routes tasks to cheaper models based on size and effort
- Delegates large repository bug fixes to a worker agent
- Checks git diffs and runs tests on worker outputs
- Runs seven hooks locally to manage task routing and compaction
- Includes a CLI tool for diagnostics, modes, and statistics
- Guards long sessions by saving notes before context compaction
- Respects model and effort ceilings from the same vendor
Use it for
- Fixing bugs in large code repositories
- Routing simple prompts to stay within the chat interface
- Delegating exploratory coding tasks to cheaper workers
- Managing long coding session contexts and compaction
Companions
Tools that articles about Token Beggar recommend using with it, each with its job.
- Claude Code — Built to work with Claude Code.
- Context7 — Provides up-to-date documentation for AI agents and code editors.
- Playwright MCP — Enables LLMs to interact with web pages using structured accessibility snapshots.
- GitHub MCP Server — Reads pull requests, issues and code search on GitHub.
- Superpowers — Brainstorm, plan and execute skills for bigger tasks.
- gstack — Skill bundle for planning, review and shipping.
Alternatives to Token Beggar
- mcp-router — Reduces Claude Code tokens by 13% to 44% using tool search and JSON shaping
- jevgrep — Reduces coding agent costs by providing targeted, relevant repository context.
- pxpipe — Cuts end-to-end token bills by compressing bulky context into compact PNGs.
- Claude Code Router — It centralizes control of AI agent requests with intelligent routing and observability.
- Effortless — It selects the reasoning effort for every prompt and lowers costs.
- Oh My ClaudeCode — Orchestrates parallel Claude, Gemini, and Codex CLI workers.
- CLI Dispatch — Runs alternative AI models as workers from inside Claude Code.
- Claude Plugin — Builds and publishes web apps directly inside Claude Code with built-in hosting.
- Edgee Mod for Claude Code — Manage models and track token cost savings inside your terminal conversation.
- Cost X-Ray — It prices usage below the tool level by analyzing raw API request bytes.
- Build with Claude — Search over 28,500 plugins, 6,900 skills, and 6,600 MCP servers.
- Lcc — Cuts token counts significantly without losing exact session details.
All Token Beggar alternatives