TokenSaver

saving tokens since May 2026

7,337,345
total tokens saved
2,629 tool runs · 58% average reduction · ≈ $22.01 saved at Sonnet input rate ($3 per 1M tokens)· 12 installs reporting
Energy and carbon dioxide you didn't have to spend 7.34 kilowatt-hours · 2.93 kilograms of carbon dioxide
illustrative — roughly 1 kilowatt-hour per 1 million tokens, United States grid average 0.4 kilograms of carbon dioxide per kilowatt-hour
Sources and methodology
  • Energy (~1 kWh per million tokens) — rough estimate derived from the IEA Electricity 2024 report and a Goldman Sachs (2024) analysis of AI data-center demand. Actual inference energy varies widely by model, provider, and hardware — easily 5–10× in either direction.
  • Carbon intensity (0.4 kg CO₂ per kWh) — United States national grid average from EPA eGRID 2023 (~0.386 kg CO₂e/kWh, rounded to 0.4). The actual figure depends on where the inference provider's data centers are located and the local grid mix.
  • Dollar savings ($3 per million input tokens) — Claude Sonnet input rate from the Anthropic pricing page. Does not account for output tokens, caching, or volume discounts.

Activity

updated 00:55:16
Tool Language Runs Saved Avg reduction Reduction
OutlineCSharpFile C# 450 2,647,207
91%
91%
Focused Emitter C# 316 1,834,117
85%
85%
Focused Emitter (multi) C# 276 1,175,196
70%
70%
MinifyFile C# 805 373,481
31%
31%
TraceImplementors C# 63 289,907
64%
64%
MinifyFile C 1 239,143
36%
36%
MinifyCSharpFile C# 307 193,814
32%
32%
TraceCallers (Initial) C# 1 158,851
46%
46%
TraceCallers C# 11 132,452
86%
86%
FocusType C# 64 116,765
73%
73%
FocusCallers C# 21 96,174
41%
41%
MinifyFile Razor 60 18,666
15%
15%
MinifyFile JSON 28 12,555
28%
28%
Cache C# 4 11,708
69%
69%
OutlineCSharpFile (Initial) C# 2 10,724
57%
57%
MinifyFile JavaScript 6 7,526
57%
57%
MinifyFile Markdown 71 4,671
0%
0%
MinifyFile XML 110 3,958
2%
2%
MinifyFile X++ 1 2,017
20%
20%
Focused Emitter (multi) (Initial) C# 1 2,004
23%
23%
MinifyFile HTML 3 1,998
1%
1%
TraceImplementors (Initial) C# 1 1,344
24%
24%
MinifyFile Python 3 713
10%
10%
Focused Emitter Cache C# 12 691
20%
20%
DedupCheck C# 2 654
81%
81%
OutlineCSharpFile VB.NET 1 391
79%
79%
Focused Emitter (Initial) C# 1 199
5%
5%
Focused Emitter VB.NET 1 168
34%
34%
MinifyFile VB.NET 1 146
30%
30%
Focused Emitter (multi) Cache C# 1 57
19%
19%
MinifyFile YAML 3 28
2%
2%
MinifyFile CSS 2 20
1%
1%

Frequently asked questions

What is TokenSaver?

TokenSaver is an MCP (Model Context Protocol) server for .NET developers. It gives AI coding assistants a token-efficient view of C#, Razor, and .NET project files using the Roslyn compiler platform, so the assistant reads the same code using far fewer tokens — with no loss of logic.

How much can TokenSaver reduce token usage?

It depends on the tool and the file. A signatures-only outline of a C# file saves 70–95%, a lossless full-file minify saves 20–50%, and the focused tools save 15–40% measured against reading only the relevant code. On C# files the typical reduction is 50–95%.

Which file types and languages does it support?

Full Roslyn support for C#, Razor, VB.NET, and .NET project files (.csproj, .props, .config, .xml). Comment-strip and whitespace-collapse support for JavaScript, TypeScript, Python, HTML, CSS/SCSS/LESS, JSON, YAML, C, C++, X++, and Markdown.

Which AI tools and editors does it work with?

Visual Studio 2026 (GitHub Copilot Chat), Claude Code, VS Code Copilot, Claude Desktop, and any MCP client that communicates over stdio.

How do I install TokenSaver?

The install page has one-click install buttons, per-client configuration snippets, and upgrade and uninstall instructions.

Is TokenSaver free and open source?

Yes. TokenSaver is open source under the MIT license, the package is published on NuGet, and the source code is on GitHub.

An unhandled error has occurred. Reload 🗙

Rejoining the server...

Rejoin failed... trying again in seconds.

Failed to rejoin.
Please retry or reload the page.

The session has been paused by the server.

Failed to resume the session.
Please retry or reload the page.