rsvp-reading

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

RSVP Reading Skill

RSVP阅读技能

Rapid Serial Visual Presentation (RSVP) is a speed reading method popularized by tools like Spritz. Words are flashed one at a time in a fixed position, while one letter per word is highlighted (typically in red) as an eye anchor — the Optimal Recognition Point (ORP). This eliminates horizontal eye movement, the primary bottleneck of traditional reading, enabling speeds of 200-600+ WPM with solid comprehension.

This skill converts any document into an RSVP token stream: each word paired with its ORP index and a calibrated display delay based on your target WPM.

Full architecture:
references/architecture.md
Acceptance criteria:
references/acceptance-criteria.md
Fallback tree:
references/fallback-tree.md
Token stream schema:
references/token-stream-schema.md

快速序列视觉呈现（Rapid Serial Visual Presentation, RSVP） 是由Spritz这类工具推广的速读方法。单词会在固定位置逐个闪现，每个单词中有一个字母（通常为红色）会被高亮作为视觉锚点，也就是最佳识别点（Optimal Recognition Point, ORP）。这种方式消除了传统阅读中水平眼动这一主要瓶颈，可以让用户在保证良好理解度的前提下达到200-600+ WPM的阅读速度。

本技能可以将任意文档转换为RSVP令牌流：每个单词都对应其ORP索引，以及基于你目标WPM校准的显示延迟。

完整架构：
references/architecture.md
验收标准：
references/acceptance-criteria.md
降级方案树：
references/fallback-tree.md
令牌流schema：
references/token-stream-schema.md

Trigger Conditions

触发条件

Invoke this skill when the user:

Says "speed read [file]", "RSVP [file]", or "read [file] at [N] WPM"
Uploads or references a document and asks to "read it fast"
Requests a token stream or reading session from a document

当用户出现以下行为时调用本技能：

说出"speed read [文件]"、"RSVP [文件]"或者"以[N] WPM的速度阅读[文件]"
上传或引用某份文档并要求"快速阅读"
请求从文档生成令牌流或阅读会话

Discovery Phase

信息收集阶段

Before executing, collect:

Input file path - What file should be parsed? (
```
.txt
```
,
```
.md
```
,
```
.pdf
```
,
```
.docx
```
)
WPM - Reading speed in words-per-minute. Default:
```
300
```
. Range:
```
100-1000
```
.
Output format - Where to save the token stream JSON? Default:
```
./rsvp_output.json
```

If any are missing, ask for them before proceeding.

执行前需要收集以下信息：

输入文件路径 - 需要解析的文件路径？（支持
```
.txt
```
、
```
.md
```
、
```
.pdf
```
、
```
.docx
```
格式）
WPM - 阅读速度，单位为词/分钟。默认值：
```
300
```
，取值范围：
```
100-1000
```
。
输出格式 - 令牌流JSON的保存路径。默认值：
```
./rsvp_output.json
```

如果有任何信息缺失，请先向用户询问再继续执行。

Execution

执行步骤

Step 1: Parse the Document

步骤1：解析文档

bash

python3 plugins/rsvp-speed-reader/skills/rsvp-reading/scripts/parse_document.py \
  --input <file_path> \
  --output /tmp/rsvp_words.json

bash

python3 plugins/rsvp-speed-reader/skills/rsvp-reading/scripts/parse_document.py \
  --input <file_path> \
  --output /tmp/rsvp_words.json

Step 2: Generate Token Stream

步骤2：生成令牌流

bash

python3 plugins/rsvp-speed-reader/skills/rsvp-reading/scripts/orp_engine.py \
  --input /tmp/rsvp_words.json \
  --wpm <wpm> \
  --output <output_path>

bash

python3 plugins/rsvp-speed-reader/skills/rsvp-reading/scripts/orp_engine.py \
  --input /tmp/rsvp_words.json \
  --wpm <wpm> \
  --output <output_path>

Step 3: Confirm Output

步骤3：确认输出

Report to the user:

Total word count
Estimated reading time at the chosen WPM
Output file path
Preview of first 5 tokens

向用户反馈以下信息：

总词数
所选WPM下的预计阅读时长
输出文件路径
前5个令牌的预览

Output Format

输出格式

Each token in the stream follows the schema in

references/token-stream-schema.md

json

{"w": "Hello", "orp": 1, "delay_ms": 200, "is_sentence_end": false, "is_para_end": false}

流中的每个令牌都遵循

references/token-stream-schema.md

中的schema：

json

{"w": "Hello", "orp": 1, "delay_ms": 200, "is_sentence_end": false, "is_para_end": false}

Confirmation Gate

确认门槛

Before generating for files > 50,000 words, display:

This document contains ~{word_count} words.
At {wpm} WPM this will take ~{minutes} minutes to read.
Generating token stream (~{token_count} tokens) to {output_path}.
Proceed? [yes/no]

当为超过50,000词的文件生成令牌流前，需要显示：

This document contains ~{word_count} words.
At {wpm} WPM this will take ~{minutes} minutes to read.
Generating token stream (~{token_count} tokens) to {output_path}.
Proceed? [yes/no]

Next Actions

后续操作

After successful generation, offer:

Open the reading session with the
```
rsvp-comprehension-agent
```
Adjust WPM and regenerate
Parse a different document

成功生成后，向用户提供以下可选操作：

调用
```
rsvp-comprehension-agent
```
开启阅读会话
调整WPM后重新生成
解析其他文档