Cut your AI coding-agent token bill

Smarter tools for Claude Code, Codex, Copilot, Cursor & OpenClaw. Fewer calls, smaller context, same model — your sessions stop ballooning.

Start 14-day free trial See how it works

No card required · 2-minute setup · your code never leaves your machine

30–50%

of an agent's tokens go to re-reading code it has already seen this session.

Every round-trip

re-ingests the whole transcript — each extra tool call makes every later call more expensive.

1 failed edit

an exact-match miss on whitespace or quotes wastes the entire turn — then the agent re-reads and retries.

How it works

12 calls vs 2

The same everyday task — find a function across files, change it everywhere, verify nothing broke.

Vanilla agent ~12 tool calls
  • 1Glob **/*.ts
  • 2Grep "renderInvoice"
  • 3Read invoice.ts (2,400 lines)
  • 4Read helpers.ts
  • 5Read types.ts
  • 6Edit invoice.ts ✕ whitespace mismatch
  • 7Read invoice.ts again, to re-anchor
  • 8Edit invoice.ts retry
  • 9Edit helpers.ts ✕ stale snippet
  • 10Read helpers.ts again
  • 11Edit helpers.ts retry
  • 12Read both files to verify
With Tokeasy 2 tool calls
  • ✓code_search "renderInvoice" — ranked matches, context, and the import graph in one call
  • ✓code_edit edits[] — batched across files, fuzzy-matched, syntax-validated before it lands

Fewer calls → smaller context → savings compound every turn.

Tools

Five tools that replace your agent's chattiest loops

Drop-in replacements for the glob-grep-read-edit cycle that burns most of your context window.

code_search

One call, not five

Glob + grep + read + imported-by in a single ranked call. Your agent gets matches, context lines, and the import graph in one round-trip.

code_read

Structure, not dumps

AST-aware truncation: long function bodies stubbed, imports, types and signatures kept. The agent sees the file's shape without ingesting 2,000 lines.

code_edit

Edits that land

Batched multi-file edits with fuzzy matching that tolerates whitespace and quote drift — plus post-edit TS/JSON/YAML syntax validation so broken edits never ship.

sql_query

Databases without the schema paste

Schema introspection, dialect auto-fix, and foreign-key graphs — the agent explores your database without dumping DDL into context.

code_savings

An honest ledger

Calls saved are counted. Bytes returned are measured. Bytes saved are estimates — and labeled that way. No made-up percentages.

MCP + OpenClaw

Works everywhere

An MCP server for Claude Code, Codex CLI, VS Code Copilot, Cursor and Windsurf — plus a native OpenClaw plugin. One license, all your harnesses.

Install

2-minute setup in your harness

Pick your tool, paste one snippet, done. The same license key works across all of them.

claude mcp add tokeasy -e TOKEASY_LICENSE_KEY=tk-... -- npx -y -p openclaw-plugin-tokeasy tokeasy-mcp

~/.codex/config.toml

[mcp_servers.tokeasy]
command = "npx"
args = ["-y", "-p", "openclaw-plugin-tokeasy", "tokeasy-mcp"]
env = { TOKEASY_LICENSE_KEY = "tk-..." }

.vscode/mcp.json

{
  "servers": {
    "tokeasy": {
      "command": "npx",
      "args": ["-y", "-p", "openclaw-plugin-tokeasy", "tokeasy-mcp"],
      "env": { "TOKEASY_LICENSE_KEY": "tk-..." }
    }
  }
}

.cursor/mcp.json

{
  "mcpServers": {
    "tokeasy": {
      "command": "npx",
      "args": ["-y", "-p", "openclaw-plugin-tokeasy", "tokeasy-mcp"],
      "env": { "TOKEASY_LICENSE_KEY": "tk-..." }
    }
  }
}
openclaw plugins install openclaw-plugin-tokeasy

openclaw.json

{
  "plugins": {
    "entries": {
      "tokeasy": {
        "config": { "licenseKey": "tk-..." }
      }
    }
  }
}
🔒

Runs 100% on your machine

Your code, prompts and keys never touch our servers. The only network call Tokeasy makes is the license check — it sends your key and an instance hash, nothing else.

Pricing

One plan. Every tool.

Try it free for 14 days — see if your token bill drops.

$9/mo
14-day free trial · no card required
  • All 5 tools — search, read, edit, SQL, savings ledger
  • Every harness: MCP + native OpenClaw plugin
  • Use on up to 5 devices
  • Savings ledger with honest, labeled numbers
  • Cancel anytime

FAQ

Questions, answered

How does Tokeasy save tokens?
Two mechanisms: fewer tool calls (one ranked search instead of a glob → grep → read chain) and smaller payloads (AST-truncated reads instead of full-file dumps). Because every tool call re-ingests the growing transcript, both effects compound — each call you skip makes every later call cheaper too.
Does it change my model or code quality?
No. Your agent keeps using the exact same model — Tokeasy only changes what flows through tool inputs and outputs. If anything, quality goes up: post-edit syntax validation catches broken TS/JSON/YAML edits before they land.
What data do you collect?
License checks only. Tokeasy runs entirely on your machine; the validate call sends your license key and an instance hash so we can enforce the device limit. Your code, prompts, file paths and query contents never leave your machine.
Which coding agents does it work with?
Claude Code, Codex CLI, VS Code Copilot, Cursor and Windsurf via MCP, plus a native OpenClaw plugin — and anything else that speaks MCP.
How do I see what I'm saving?
Ask your agent to run the code_savings tool. It reports a ledger where calls saved are counted, bytes returned are measured, and bytes saved are estimates — each number labeled for what it actually is.