Model Salt

com.modelsalt由 modelsalt.com 发布官方发布者v1.0.4更新于 Nov 5, 2026

Randomness from outside the model, so one prompt gives different answers that still fit the ask.

已验证Streamable HTTP可网页运行开发者工具AI 与机器学习

概览

AI 生成的概览

Model Salt 为创作类请求引入模型之外的随机性，让助手对同一提示给出不同但仍符合要求的回答。

可以做什么
  • 用 model_salt 工具原样传入请求，取回本次需要确定的开放选项
  • 带上返回的 id 和助手对探索步骤的回复再次调用，完成该步骤
  • 用 0 到 2 的强度设置（默认 1）调节结果偏离模型惯常选择的程度
  • 通过 keep 字段保留已确定的答案，只重新抽取其余开放选项
  • 对只有一个正确答案的问题（如计算、事实、修 bug）跳过该工具
适用场景
适合希望同一开放式提示多次运行得到真正不同结果的场景，例如图片、SVG、命名、故事、计划或设计。不适合事实性或单一答案的任务，此时模型会被告知跳过该工具。
运行要求
需要 Node 18 或更高版本以运行 npx 包，或使用远程端点。安装命令会把服务器加入 Claude Code、Codex、Gemini CLI 或 Kimi Code。可选：已登录的 Claude Code 或 Codex，会在你自己的套餐上执行一次简短的额外请求；若没有，模型会在自己的对话中完成该步骤并消耗更多 token。访问公共随机源需要网络，离线时回退到本机随机数。
安装前请注意
安装会修改客户端的 MCP 配置，并为 Codex 和 Kimi Code 在 ~/.agents/skills/ 下添加技能文件；卸载会移除它们。加盐后的回答更慢且更耗 token（示例中每条回答多约 3,300 个输出 token）。提示词不会发送给随机源，服务端也不存储内容，但该工具确实会改变助手的回答方式。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Model Salt，将其添加到工作区。
  2. 为需要使用其工具的对话启用该服务。

Web executable，通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体，或 SourceWeft 对话中。智能体会读取该服务器的安装指南，先向你展示发布者、端点和登录方式，经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/com.modelsalt%2Fmodel-salt/install.md，并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前，先告诉我它的发布者、端点和登录方式，等我确认。不要把 API 密钥写进命令或我会提交的文件；问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。

{
  "mcpServers": {
    "model-salt": {
      "type": "http",
      "url": "https://modelsalt.com/mcp"
    }
  }
}

README

Model Salt

Ask a language model for the same thing ten times and you get the same answer ten times. Model Salt is an MCP server that gives your model randomness from outside it, so one prompt gives ten different answers that still do exactly what you asked.

[npm] [MIT licence]

modelsalt.com · works in Claude Code, Codex, Gemini CLI and Kimi Code

[Ten runs of "Generate an SVG of a pelican riding a bicycle" on Claude Opus 5.5: the plain prompt draws the same pelican ten times, with Model Salt nine of the ten pictures are different]

Install

One command adds it to your client. Run it once.

ClientCommand
Claude Codenpx -y @modelsalt/model-salt setup claude
Codexnpx -y @modelsalt/model-salt setup codex
Gemini CLInpx -y @modelsalt/model-salt setup gemini
Kimi Codenpx -y @modelsalt/model-salt setup kimi

Start a new session in your client after it, and the model_salt tool is there.

  • What it adds: the model-salt server in your client's MCP list. For Codex and Kimi Code it also adds a short skill at ~/.agents/skills/model-salt/ that tells the model when to call the tool, and for Codex one setting on that server so the model sees the tool from the start.
  • Gemini CLI runs local servers only in folders you trust.
  • See first: add --dry-run to print what it would change and change nothing.
  • Run it twice: nothing changes the second time.

Undo

sh
npx -y @modelsalt/model-salt uninstall claude    # or codex, gemini, kimi

It takes out everything setup added. The skill stays while another client of yours still uses it. --dry-run works here too.

Use it

Ask for anything with room for choices: a picture, an SVG, a page, a name, a story, a plan. The model calls model_salt with your request, gets back what to settle this time, and then does your request as you asked it.

Want more or less of it? Say so in your prompt ("make it really different", "keep it close to the usual"). The tool takes a setting from 0 to 2. Left out, it is 1.

SettingHow often each choice keeps the model's usual pick (target)
0always
0.570% of the time
1 (default)45% of the time
1.525% of the time
2no more often than any other fitting pick

The model skips the tool for questions with one right answer (a sum, a fact, a bug fix).

Keep what you like. Like part of a result? Say what to keep in plain words ("keep the palette and layout, give me new illustrations"). Your model copies those decided answers from the earlier Model Salt reply into the tool's keep field and calls it again. Only the open choices are rolled again, and the reply starts with one line, e.g. Kept: palette = burnt orange on cream; layout = split screen, text left. That holds on later runs until you say start fresh, and "give me something new" leaves it out. In a new chat, paste the Kept: line or say it in your own words. A kept answer that no longer fits a new request is dropped, and the reply says so. Nothing is stored on our side.

What it changes

Same prompt, 10 runs each, from the examples on modelsalt.com.

"Generate an SVG of a pelican riding a bicycle." Different pictures in 10 runs:

ModelPlain promptWith Model Salt
Claude Opus 5.519
Claude Fable 5.139
GPT-6.1 Sol210
GPT-6 Astra18
Gemini 4 Argon210
Kimi K319

"Suggest a name for a neighbourhood coffee shop." Claude Opus 5.5, the first five runs:

Plain promptWith Model Salt
The Common GrindThe Footbridge
Common GroundBrew Espresso Bar
The Common GrindSnugwick Corner
The Daily Grind & GatherOld Oak's
The Common CupHilltop Coffee

"Write the first line of a short story about a lighthouse keeper." Claude Opus 5.5:

  • Plain, 5 of 5 runs: "For thirty-one years, Elias Marrow had kept the light burning for ships that never..." (Marrow, Maren or Marsh, thirty-one or forty-one years)
  • With Model Salt: "The sea has gone perfectly still, and I have never been more afraid." / "The beam swept the black water again, lighting nothing Tomas had ever loved." / "Halfway up the iron stair with a can of whale oil knocking against her knee, Agnes Merrow..."

Every run, plain and salted, is in examples/, with the prompt, the model and the answer.

What it needs

  • Node 18 or later. npx fetches the package the first time.
  • Claude Code or Codex, signed in, or neither. With one of them on the machine, each call runs one short extra request on your machine, on your own plan, with no API key: through Codex when you use Codex, and through Claude Code (Claude Opus 5.5) when you use Claude Code, Gemini CLI or Kimi Code. If that request can't run (signed out), the tool says so and your model carries on with the request as asked. With neither on the machine (a Docker container, Cursor or Copilot alone), your own model does that step in its own chat: it calls the tool twice, first with the request, then with its reply to the exploring step, and uses more of its own tokens for it.
  • Time and tokens. A salted answer takes longer and uses more tokens than a plain one. On the pelican prompt, Claude Opus 5.5 used about 3,300 more output tokens per answer.
  • Network. The server fetches random bytes from public randomness sources (quantum noise from the LfD quantum random number service, and the public drand and NIST beacons). Your prompt is not sent to them. Offline, it uses your computer's own random generator.

Other ways to run it

  • Short name: npx -y model-salt <args> runs the same package as npx -y @modelsalt/model-salt <args>.
  • As a plugin: plugin/ holds a Codex plugin, a Claude Code plugin and a Gemini CLI extension, and this repo is a plugin marketplace. From a clone:
    • Codex: codex plugin marketplace add /path/to/model-salt, then codex plugin add model-salt@model-salt
    • Claude Code: claude plugin marketplace add /path/to/model-salt, then claude plugin install model-salt@model-salt
    • Gemini CLI: gemini extensions install /path/to/model-salt/plugin
  • By hand: add a stdio MCP server that runs npx -y @modelsalt/model-salt anthropic (Claude, Gemini and Kimi models) or npx -y @modelsalt/model-salt openai (GPT models).
  • From a clone: node src/stdio.mjs anthropic in place of the npx command.

How it works

In a separate short run (with no Claude Code or Codex on the machine, in your own model's chat, over two tool calls), the model names the biggest open decisions in your request (the questions an expert would ask) and a range of answers to each, its usual one included. Randomness from outside the model picks one answer per question. Your model then does the request with those answers settled. More, and why simpler ways failed: docs/how-it-works.md.

Contributing

Found a prompt that still gives the same answer every time with Model Salt? Open an issue with the prompt, the model and a few runs. How to test a change to a family's steps: CONTRIBUTING.md.

Privacy

Model Salt runs on your machine and never sends your prompts anywhere. It only asks public random sources for random bytes. Details: Privacy policy.

Licence

  • Code: MIT, see LICENSE.
  • Example runs and pictures (examples/, the images in docs/): CC BY 4.0, credit "Model Salt", see examples/LICENSE.
  • Word list (src/words.mjs, Webster's 2nd, 1934): public domain.

Model Salt is a project by Chris Benjaminsen.

来源：README.md，提交 61b3712

工具

1

model_salt

Model Salt

Call this before any creative or open-ended request (an image, a picture prompt, a story, a poem, a name, an idea, a design, a plan), so the result is varied instead of your usual answer and still exactly what was asked. First call: send the user's request word for word in `request` (and `randomness` only if the user asked for more or less variety). It returns an exploring step and an `id`. Do that step silently, without showing it to the user, and call this tool again with the `id` and your whole reply to the step in `explored`. The second call returns what was decided. Then make the result itself in the same reply: for a request for an image, generate the image with your image generation tool and show it (a written prompt is not the result); for writing, write it. The result shows every decided answer.

输入参数idkeeprequestexploredrandomness

版本历史

3
  1. v1.0.4最新Nov 5, 2026
  2. v1.0.2Nov 3, 2026
  3. v1.0.1Oct 29, 2026

评分

已满足适用检查项的 100%

该评分衡量此收录条目的完整程度，依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价，也不表示服务器安全或当前可用：可连通性是带日期的检查结果，评分也不包含安装量或用户评价。

  • 近 30 天内可连通必需25 分
  • 已声明认证方式必需10 分
  • 已列出工具必需15 分
  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分