Token usage and cost tracking tools for coding agents
Trackers for token usage and spend across coding agents. Of the 391 tools on our writing code page, 10 fit this narrower job. By pricing: 8 open source, 1 source-available, 1 paid, usage-based or enterprise. The most-starred open repos here are rtk (84,058 stars), Trail Code (10,211 stars), Mission Control (6,358 stars). Most come from the coding (6), agent infra (2), MCP servers (1) categories.
Updated
rtk
A CLI proxy that sits between coding agents and your shell and trims noisy command output before it reaches the model, cutting token use on everyday commands like tests, builds and git. Free and open source, with an RTK Pro tier for teams.
Added
Trail Code
Context layer that makes Claude Code, Cursor, Codex and other coding agents faster and cheaper on large codebases.
Added
Mission Control
Self-hosted control plane for dispatching tasks to agents, reviewing runs and tracking spend.
Added
Tokscale
Tracks token usage across AI coding agents from your terminal.
Added
Weave Router
An open-source LLM router for Claude Code, Codex and other coding agents that sends each request to the cheapest model that still meets your quality bar, cutting model spend.
Added
abtop
htop-style monitor for Claude Code and Codex sessions and token use.
Added
jCodeMunch MCP
MCP server for symbol-level code exploration that cuts token use.
Added
Jevgrep
CLI that finds relevant repo files and excerpts from natural-language questions for coding agents; claims ~40% cheaper agent runs.
Added
HiRoute
HiRoute, from the Higress team, is an open-source local model router for long-running agents like Codex and Claude Code: it hands routine stages to cheaper models, saves your strongest model for hard steps and keeps a record of each decision.
Added
Empyre
Empyre turns a plain-language business brief into a running software company: eight AI agents (CEO, CTO, CMO, CFO and more) build it, deploy it to its own HTTPS address and keep operating it, shipping code, marketing and support with spend approval; plans $49, $149 and $299 a month.
Added
GitHub repos (9)
| Repo | GitHub stars | Licence | Last push |
|---|---|---|---|
| rtk ↗ | 84,058 | — | 2026-10-20 |
| Trail Code ↗ | 10,211 | MIT | 2026-10-21 |
| Mission Control ↗ | 6,358 | MIT | 2026-09-22 |
| Tokscale ↗ | 5,695 | MIT | 2026-10-20 |
| Weave Router ↗ | 5,595 | — | 2026-10-21 |
| abtop ↗ | 3,752 | MIT | 2026-09-14 |
| jCodeMunch MCP ↗ | 2,776 | jCodeMunch-MCP Dual-Use License (free for non-commercial use) | 2026-10-21 |
| Jevgrep ↗ | 2,675 | MIT | 2026-10-21 |
| HiRoute ↗ | 278 | — | 2026-10-21 |
9 repos. Stars and licence come from each repo, checked when the listing was last verified.
FAQ
What are the best AI tools for writing code focused on token usage and cost tracking?
We list 10. The first three by GitHub stars and featured status are rtk: A CLI proxy that sits between coding agents and your shell and trims noisy command output before it reaches the model, cutting token use on everyday commands like tests, builds and git. Free and open source, with an RTK Pro tier for teams.; Trail Code: Context layer that makes Claude Code, Cursor, Codex and other coding agents faster and cheaper on large codebases.; Mission Control: Self-hosted control plane for dispatching tasks to agents, reviewing runs and tracking spend.
Are there free or open-source options?
Yes: 8 open source, 1 source-available, 1 paid, usage-based or enterprise. Pricing is as each maker states it.
More from writing code
Tools are grouped by what each says it does. Missing one? Submit it free.