MCP服务器在真实编码基准测试中，活跃token数-80%，墙面时间-83%，而普通时效为78.3%，达到97.9%（188/192）。结构代码导航+持久内存引擎。
概览
一个MCP服务器。一个配置文件。为AI编码代理提供结构代码导航、持久内存和Bash输出压缩。 ——项目场地 + 基准着陆 ——基准来源 + 装置 使用优化配置文件（单一环境变量）重现。参见基准总结。 真实工作台对比7天的转录（1130个Bash输出）驱动了此次发布。累计节省现为（19.3%匹配率，平均压缩率68.9%），而v4.2.0每周约1.2千美元。 - 固定 pytest 正则表达式：现在匹配 python3 -m pytest、uv 运行 pytest、venv 前缀表单、poetry/hatch/pdm/rye 运行 pytest。- 再用 5 个 git 压缩器：fetch、checkout、branch、worktree list、stash list。- 4个 gh 压缩器：gh repo 视图、gh pr 视图、gh issue 视图、gh pr diff。- grep、find、cat 压缩器。按文件组命中，去除常见前缀，头尾截断。夹具节省 83-96%。- 复合命令分流器：cd /root/foo & git 状态现在通过选择 &&/;链中最后一个有意义的段来压缩。
概述（AI 翻译）
一个MCP服务器。一个配置文件。在tsbench上97.9%，tokens-80%。 结构代码导航、持久内存和 Bash 命令重写，用于 AI 编码代理。
-亮绿色）
mibayy.github.io/token-savior ——项目现场 + 基准着陆 基准数据来源 + 赛程：目前未公布（见下文转载比分*）
v4.23 31秒，测量时间为2026-10-07 ·观看完整视频（MP4，开声）
基准测试——96个真实编码任务（Claude Opus 4.7,2026年5月）
| 普通克劳德密码 | 与象征救世主合作 | |
|---|---|---|
| 得分141 / 180（78.3%） | 188 / 192（97.9%） | |
| 活跃标记/任务 | 17 221 | 3 395（-80%） |
| 墙壁时间/任务 | 110.6秒 | 18.9 s（-83%） |
用optimized配置文件（单一环境变量）来复制。束束
这些编号的生成如下描述;其存储库不公开于
因此，应按报告数据取，而非独立验证。
重新测量于2026-08-09发布，现已撤回 2026-08-10. 报告称，使用小型替换线束节省了新代币。
本页由 AI 翻译，技术细节请以原文为准。
README
以下为原文 README 节选，完整内容请查看原文。
Token Savior
One MCP server. One profile. 97.9% on tsbench at -80% tokens. Structural code navigation, persistent memory, and Bash command rewriting for AI coding agents.
-brightgreen){kind=icon}
mibayy.github.io/token-savior – project site + benchmark landing Benchmark source + fixtures: not currently published (see Reproducing the score below)
v4.23 in 31 seconds, measured on 2026-10-07 · watch the full video (MP4, sound on)
Benchmark – 96 real coding tasks (Claude Opus 4.7, May 2026)
| Plain Claude Code | With Token Savior | |
|---|---|---|
| Score | 141 / 180 (78.3%) | 188 / 192 (97.9%) |
| Active tokens / task | 17 221 | 3 395 (-80%) |
| Wall time / task | 110.6 s | 18.9 s (-83%) |
Reproduces with the optimized profile (single env var). The harness that
produced these numbers is described below; its repository is not public at the
moment, so take the figures as reported rather than as independently verifiable.
A re-measurement was published here on 2026-08-09 and has been withdrawn on
2026-08-10. It reported new-token savings from a small replacement harness.
Those numbers did not measure this server at all: across 143 benchmark
sessions, exactly one called a Token Savior tool. The client running the
harness had MCP deferred-tool loading enabled, so all 18 tools sat behind a
ToolSearch lookup instead of appearing in the model’s manifest. The model
never saw them and fell back to Grep and Read — 66 greps, 30 reads, one
MCP call. What varied between the “profiles” was the size of the cached
prefix, not what the agent did.
The lesson is worth more than the numbers were: a benchmark of an MCP server must assert that its tools were actually called. Ours did not, so it happily compared two identical agents. That assertion now exists in the harness.
安装
uvx token-savior-recall配置
{
"mcpServers": {
"token-savior-recall": {
"command": "/path/to/venv/bin/token-savior",
"alwaysLoad": true,
"env": {
"WORKSPACE_ROOTS": "/path/to/project1,/path/to/project2",
"TOKEN_SAVIOR_CLIENT": "claude-code",
"TOKEN_SAVIOR_PROFILE": "optimized"
}
}
}
}