classifier.dev
Sort up to 1,000 texts into your own labels with a calibrated confidence per answer. No API key.
我该使用它吗
质量与安全性
A
基于对工具定义和协议合规性的自动分析。
上下文开销
~3,152token 数（工具定义）
~6.1 KB典型响应大小
对注意力有显著影响（占 128k 上下文窗口的 2.46%）
这是每次将服务器的工具加载到模型上下文窗口时所消耗的大致 token 数。数值越高，可用于其他任务的注意力就越少。
安装
一键安装
将以下内容添加到你的 `claude_desktop_config.json` 文件中：
{
"mcpServers": {
"classifier": {
"url": "https://classifier.dev/mcp"
}
}
}远程端点
https://classifier.dev/mcpstreamable-http它能做什么
工具清单
工具（5）
🟢 只读🟡 写入🔴 删除⚪ 未知
🟢classify_texts(url, include, inputs, labels, instructions, ...)
Sort up to 1,000 texts into exactly one of your own labels each, with confidence per answer. Use this when you have many items to triage, route, filter or bucket and do not want to read them all: search results before opening them, tickets, log lines, changed files, feedback. Do not use it for fewer than about five items you can already see — just decide. For default Jev, confidence is calibrated (answers >= 0.9 are right ~82-92% of the time; < 0.5 about 30-60%); these measurements do not apply to experimental Laya. so act on the sure ones and look at the rest yourself, or pass tier "smart" to have the unsure ones re-asked of a reasoning model.
输入模式
{
"type": "object",
"properties": {
"url": {
"type": "string",
"pattern": "^https?://",
"maxLength": 8192,
"description": "Scrape one public URL instead of inputs/items. Requires funded workspace access. Context.dev costs $0.0022 per billed attempt plus classification; long articles need Fast. Errors disclose retained charges. No automatic scrape retries."
},
"include": {
"type": "array",
"items": {
"type": "string",
"enum": [
"markdown",
"html"
]
},
"description": "With url: return article.markdown and/or article.html at no extra scrape cost. Omit for compact output."
},
"inputs": {
"type": "array",
"items": {
"type": "string"
},
"minItems": 1,
"maxItems": 1000,
"description": "1 to 1,000 texts to classify. Results come back in the same order."
},
"labels": {
"type": "array",
"items": {
"type": "string"
},
"minItems": 2,
"maxItems": 100,
"description": "2 to 100 category names. Descriptive names classify better: \"urgent bug\" beats \"p0\". Add a label like \"none of these\" when none-of-the-above is a real outcome."
},
"instructions": {
"type": "string",
"description": "Optional extra criteria, e.g. \"judge only the service, ignore the food\"."
},
"tier": {
"type": "string",
"enum": [
"fast",
"smart"
],
"description": "fast (default) or smart, which re-asks answers under 0.7 confidence of a reasoning model (slower, single-label only). Independent of the Laya processing lane."
},
"model": {
"type": "string",
"enum": [
"jev",
"laya",
"kev",
"chunklaya"
],
"description": "Jev is default. Default/explicit jev inputs over 32,000 characters use paid Fast-only long context: up to 250,000 original cl100k_base context tokens total, 20 documents, 32 decisions and a 1 MB body. Requires paid workspace balance or active paid subscription, not signup credit. $0.084/M original context tokens counted once across inputs, independent of dimensions. Final Jev uses selected evidence; eligible chunks may be omitted, disclosed in usage.long_context. No evidence returns 422 long_context_no_evidence without charge. Explicit 'chunklaya' retains legacy opt-in (4,000,000 characters/input, 20 inputs, subject to body limit). 'laya' (512-token context) and 'kev' (8K context) remain experimental alternatives."
},
"processing": {
"type": "string",
"enum": [
"fast",
"bulk"
],
"description": "Optional. Implies Laya if model is omitted; has no effect with explicit Jev. With Laya, omit to select fast for one decision or bulk for batches automatically. Explicit fast accepts one decision. Shared capacity limits can return 429."
}
},
"required": [
"labels"
],
"oneOf": [
{
"required": [
"inputs"
],
"not": {
"required": [
"url"
]
}
},
{
"required": [
"url"
],
"not": {
"required": [
"inputs"
]
}
}
],
"additionalProperties": false
}输出模式
{
"type": "object",
"properties": {
"tier": {
"type": "string"
},
"model": {
"type": "string"
},
"results": {
"type": "array",
"description": "One per input, in input order.",
"items": {
"type": "object",
"properties": {
"label": {
"type": "string"
},
"confidence": {
"type": [
"number",
"null"
],
"description": "0-1; calibration measurements cover Jev, not experimental Laya. Null when the provider returns no score or the smart tier replaces the scored answer."
},
"scores": {
"type": [
"object",
"null"
],
"additionalProperties": {
"type": "number"
},
"description": "Model preference per supplied label; sums to 1. Does not validate the input or guarantee correctness."
},
"escalated": {
"type": "boolean",
"description": "Smart tier only: this answer was re-asked of the reasoning model."
}
},
"required": [
"label"
]
}
},
"usage": {
"type": "object",
"properties": {
"classifications": {
"type": "integer"
},
"escalated": {
"type": "integer"
},
"ms": {
"type": "integer"
}
}
}
},
"required": [
"results"
]
}🟢classify_dimensions(url, include, items, dimensions, instructions, ...)
Classify each text by several named dimensions, such as team, urgency and kind, in one request. Returns a label, confidence, scores and model for each field. At most 1,000 item × dimension decisions; every field counts toward the quota. Use per-dimension instructions to define ambiguous categories.
输入模式
{
"type": "object",
"properties": {
"url": {
"type": "string",
"pattern": "^https?://",
"maxLength": 8192,
"description": "Scrape one public URL instead of inputs/items. Requires funded workspace access. Context.dev costs $0.0022 per billed attempt plus classification; long articles need Fast. Errors disclose retained charges. No automatic scrape retries."
},
"include": {
"type": "array",
"items": {
"type": "string",
"enum": [
"markdown",
"html"
]
},
"description": "With url: return article.markdown and/or article.html at no extra scrape cost. Omit for compact output."
},
"items": {
"type": "array",
"items": {
"type": "string"
},
"minItems": 1,
"maxItems": 1000,
"description": "1 to 1,000 texts to classify. Results come back in the same order."
},
"dimensions": {
"type": "object",
"minProperties": 1,
"maxProperties": 20,
"propertyNames": {
"minLength": 1,
"maxLength": 64,
"pattern": "\\S"
},
"description": "Named dimensions. Each is a label array or {labels, instructions}. At most 1,000 item × dimension decisions; definitions at most 16,000 characters combined.",
"additionalProperties": {
"oneOf": [
{
"type": "array",
"minItems": 2,
"maxItems": 100,
"uniqueItems": true,
"items": {
"type": "string",
"minLength": 1,
"maxLength": 200,
"pattern": "\\S"
}
},
{
"type": "object",
"required": [
"labels"
],
"additionalProperties": false,
"properties": {
"labels": {
"type": "array",
"minItems": 2,
"maxItems": 100,
"uniqueItems": true,
"items": {
"type": "string",
"minLength": 1,
"maxLength": 200,
"pattern": "\\S"
}
},
"instructions": {
"type": "string",
"maxLength": 4000
}
}
}
]
}
},
"instructions": {
"type": "string",
"description": "Optional extra criteria, e.g. \"judge only the service, ignore the food\".",
"maxLength": 4000
},
"tier": {
"type": "string",
"enum": [
"fast",
"smart"
],
"description": "fast (default) or smart, which re-asks answers under 0.7 confidence of a reasoning model (slower, single-label only). Independent of the Laya processing lane."
},
"model": {
"type": "string",
"enum": [
"jev",
"laya",
"kev",
"chunklaya"
]
},
"processing": {
"type": "string",
"enum": [
"fast",
"bulk"
],
"description": "Optional. Implies Laya if model is omitted; has no effect with explicit Jev. Omit for automatic fast/bulk selection based on item × dimension decisions."
}
},
"required": [
"dimensions"
],
"oneOf": [
{
"required": [
"items"
],
"not": {
"required": [
"url"
]
}
},
{
"required": [
"url"
],
"not": {
"required": [
"items"
]
}
}
],
"additionalProperties": false
}🟢classify_multi_label(url, include, inputs, labels, instructions, ...)
Like classify_texts, but each text gets every label that applies (possibly none), with an independent 0-1 score per label. Use this for tagging — topics of an article, components touched by a ticket — where one answer is not enough. Set max_labels to cap how many come back per text. Labels scoring >= 0.7 are kept.
输入模式
{
"type": "object",
"properties": {
"url": {
"type": "string",
"pattern": "^https?://",
"maxLength": 8192,
"description": "Scrape one public URL instead of inputs/items. Requires funded workspace access. Context.dev costs $0.0022 per billed attempt plus classification; long articles need Fast. Errors disclose retained charges. No automatic scrape retries."
},
"include": {
"type": "array",
"items": {
"type": "string",
"enum": [
"markdown",
"html"
]
},
"description": "With url: return article.markdown and/or article.html at no extra scrape cost. Omit for compact output."
},
"inputs": {
"type": "array",
"items": {
"type": "string"
},
"minItems": 1,
"maxItems": 1000,
"description": "1 to 1,000 texts to classify. Results come back in the same order."
},
"labels": {
"type": "array",
"items": {
"type": "string"
},
"minItems": 2,
"maxItems": 100,
"description": "2 to 100 category names. Descriptive names classify better: \"urgent bug\" beats \"p0\". Add a label like \"none of these\" when none-of-the-above is a real outcome."
},
"instructions": {
"type": "string",
"description": "Optional extra criteria, e.g. \"judge only the service, ignore the food\"."
},
"max_labels": {
"type": "integer",
"minimum": 1,
"maximum": 100,
"description": "At most this many labels per text, most likely first."
},
"model": {
"type": "string",
"enum": [
"jev",
"laya",
"kev",
"chunklaya"
]
},
"processing": {
"type": "string",
"enum": [
"fast",
"bulk"
],
"description": "Optional. Implies Laya if model is omitted; has no effect with explicit Jev. Omit to select fast for up to four labels on one text, or bulk for larger work automatically."
}
},
"required": [
"labels"
],
"oneOf": [
{
"required": [
"inputs"
],
"not": {
"required": [
"url"
]
}
},
{
"required": [
"url"
],
"not": {
"required": [
"inputs"
]
}
}
],
"additionalProperties": false
}输出模式
{
"type": "object",
"properties": {
"results": {
"type": "array",
"items": {
"type": "object",
"properties": {
"labels": {
"type": "array",
"items": {
"type": "string"
},
"description": "Every label scoring >= 0.7, most likely first. May be empty."
},
"scores": {
"type": [
"object",
"null"
],
"additionalProperties": {
"type": "number"
},
"description": "Independent model preference per supplied label. Does not validate the input or guarantee correctness."
}
},
"required": [
"labels"
]
}
},
"usage": {
"type": "object",
"properties": {
"classifications": {
"type": "integer"
},
"ms": {
"type": "integer"
}
}
}
},
"required": [
"results"
]
}🟢count_labels(inputs, labels, instructions, unsure_below)
Classify up to 1,000 texts and return only a histogram: how many landed on each label, and how many the model was unsure about. Use this when you want the shape of a corpus — what share of feedback is bugs vs praise, how many search results are relevant — without pulling a thousand individual answers into context. Use classify_texts when you need the answer per item.
输入模式
{
"type": "object",
"properties": {
"inputs": {
"type": "array",
"items": {
"type": "string"
},
"minItems": 1,
"maxItems": 1000,
"description": "1 to 1,000 texts to classify. Results come back in the same order."
},
"labels": {
"type": "array",
"items": {
"type": "string"
},
"minItems": 2,
"maxItems": 100,
"description": "2 to 100 category names. Descriptive names classify better: \"urgent bug\" beats \"p0\". Add a label like \"none of these\" when none-of-the-above is a real outcome."
},
"instructions": {
"type": "string",
"description": "Optional extra criteria, e.g. \"judge only the service, ignore the food\"."
},
"unsure_below": {
"type": "number",
"minimum": 0,
"maximum": 1,
"default": 0.7,
"description": "Answers with confidence under this count as unsure."
}
},
"required": [
"inputs",
"labels"
],
"additionalProperties": false
}输出模式
{
"type": "object",
"properties": {
"total": {
"type": "integer"
},
"counts": {
"type": "object",
"additionalProperties": {
"type": "integer"
},
"description": "Label -> how many texts, every label present."
},
"unsure": {
"type": "integer",
"description": "How many answers fell under unsure_below."
},
"unsure_below": {
"type": "number"
}
},
"required": [
"total",
"counts",
"unsure"
]
}🟢review_uncertain(inputs, labels, instructions, below)
Classify up to 1,000 texts and return only the ones whose confidence fell under a threshold (default 0.7), each with its two most likely labels. Use this after a bulk classification to decide which items deserve your own attention: the confident answers can be trusted, these are the ones to read. Returns the index of each item so you can map back to your list.
输入模式
{
"type": "object",
"properties": {
"inputs": {
"type": "array",
"items": {
"type": "string"
},
"minItems": 1,
"maxItems": 1000,
"description": "1 to 1,000 texts to classify. Results come back in the same order."
},
"labels": {
"type": "array",
"items": {
"type": "string"
},
"minItems": 2,
"maxItems": 100,
"description": "2 to 100 category names. Descriptive names classify better: \"urgent bug\" beats \"p0\". Add a label like \"none of these\" when none-of-the-above is a real outcome."
},
"instructions": {
"type": "string",
"description": "Optional extra criteria, e.g. \"judge only the service, ignore the food\"."
},
"below": {
"type": "number",
"minimum": 0,
"maximum": 1,
"default": 0.7,
"description": "Return items with confidence under this."
}
},
"required": [
"inputs",
"labels"
],
"additionalProperties": false
}输出模式
{
"type": "object",
"properties": {
"total": {
"type": "integer"
},
"below": {
"type": "number"
},
"uncertain": {
"type": "array",
"items": {
"type": "object",
"properties": {
"index": {
"type": "integer",
"description": "Position in the inputs you sent."
},
"text": {
"type": "string"
},
"label": {
"type": "string",
"description": "The model's best guess."
},
"confidence": {
"type": [
"number",
"null"
]
},
"runner_up": {
"type": [
"string",
"null"
],
"description": "The second most likely label."
}
},
"required": [
"index",
"text",
"label"
]
}
}
},
"required": [
"total",
"uncertain"
]
}社区
证据
最近观测
已验证未记录版本5 个工具
已验证未记录版本5 个工具