parallel-ai-search

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Parallel AI Search (CLI Master)

Parallel AI 搜索(CLI 主控技能)

This is a single “master” skill that replaces the earlier Node-script-based version of
parallel-ai-search
.
It routes to the right
parallel-cli
capability for the task:
  • Search: quick web lookup with citations (
    parallel-cli search
    )
  • Extract: turn URLs (including PDFs and JS-heavy pages) into clean, LLM-ready text (
    parallel-cli extract
    )
  • Deep research: multi-source reports with processor tiers (
    parallel-cli research ...
    )
  • Enrich: add web-sourced columns to CSV/JSON (
    parallel-cli enrich ...
    )
  • FindAll: discover entities from the web with optional enrichments (
    parallel-cli findall ...
    )
  • Monitor: track web changes on a cadence, optionally via webhook (
    parallel-cli monitor ...
    )
这是一项单一的“主控”技能,取代了早期基于Node脚本的
parallel-ai-search
版本。
它会根据任务需求路由到
parallel-cli
对应的功能:
  • 搜索:带引用的快速网页查询(
    parallel-cli search
  • 提取:将URL(包括PDF和JS渲染页面)转换为适合LLM处理的整洁文本(
    parallel-cli extract
  • 深度研究:带处理器层级的多来源报告(
    parallel-cli research ...
  • 增强:为CSV/JSON添加网页来源的列(
    parallel-cli enrich ...
  • FindAll:从网页中发现实体并可选择进行增强(
    parallel-cli findall ...
  • 监控:按周期跟踪网页变化,可通过webhook推送(
    parallel-cli monitor ...

Routing rules (pick ONE)

路由规则(选择其中一项)

Choose the smallest / cheapest action that solves the user’s request:
  1. Extract — if the user gives one or more URLs or says “read/summarise this page”, “extract”, “quote”, “pull the content”, “what does this page say”.
  2. Deep research — ONLY if the user explicitly asks for deep, exhaustive, comprehensive, thorough investigation, or a multi-source “report”.
  3. Enrich — if the user provides a list/table (CSV/JSON/inline objects) and wants new columns like CEO, revenue, funding, contact info, etc.
  4. FindAll — if the user wants you to discover many entities (companies/people/venues/etc.) that match criteria.
  5. Monitor — if the user wants ongoing tracking (“alert me”, “track changes”, “monitor this weekly”) rather than a one-off answer.
  6. Search — default for everything else that needs current web info or citations.
Optional manual prefixes if the user invoked this skill directly:
  • search: ...
  • extract: ...
  • research: ...
  • enrich: ...
  • findall: ...
  • monitor: ...
If a prefix is present, honour it.
选择能解决用户需求的最小/成本最低的操作:
  1. 提取 — 如果用户提供一个或多个URL,或者提到“读取/总结这个页面”、“提取”、“引用”、“获取内容”、“这个页面讲了什么”。
  2. 深度研究 — 仅当用户明确要求深度详尽全面彻底调查,或者需要多来源“报告”时使用。
  3. 增强 — 如果用户提供列表/表格(CSV/JSON/内联对象),并希望添加CEO、收入、融资、联系信息等新列。
  4. FindAll — 如果用户希望你发现符合条件的大量实体(公司/人物/场所等)。
  5. 监控 — 如果用户需要持续跟踪(“提醒我”、“跟踪变化”、“每周监控这个内容”)而非一次性答案。
  6. 搜索 — 所有其他需要最新网页信息或引用的默认选项。
如果用户直接调用此技能,可使用可选的手动前缀:
  • search: ...
  • extract: ...
  • research: ...
  • enrich: ...
  • findall: ...
  • monitor: ...
如果存在前缀,请遵循该前缀执行操作。

Setup and authentication (only when needed)

设置与身份验证(仅在需要时进行)

Before running any Parallel command, ensure auth works:
bash
parallel-cli auth
If
parallel-cli
is missing, install it:
bash
curl -fsSL https://parallel.ai/install.sh | bash
If you cannot use the install script, use pipx:
bash
pipx install "parallel-web-tools[cli]"
pipx ensurepath
Then authenticate (choose one):
bash
undefined
运行任何Parallel命令前,请确保身份验证正常:
bash
parallel-cli auth
如果缺少
parallel-cli
,请安装:
bash
curl -fsSL https://parallel.ai/install.sh | bash
如果无法使用安装脚本,可使用pipx:
bash
pipx install "parallel-web-tools[cli]"
pipx ensurepath
然后进行身份验证(选择其中一种方式):
bash
undefined

Interactive OAuth (opens browser)

交互式OAuth(打开浏览器)

parallel-cli login
parallel-cli login

Headless / SSH / CI

无界面/SSH/CI环境

parallel-cli login --device
parallel-cli login --device

Or environment variable

或使用环境变量

export PARALLEL_API_KEY="your_api_key"
undefined
export PARALLEL_API_KEY="your_api_key"
undefined

Output & citation rules

输出与引用规则

  • Always cite web-sourced facts with inline markdown links:
    [Source Title](https://...)
    .
  • End with a Sources list whenever you used Search/Extract/Research output.
  • Prefer official/primary sources when available.
  • For long outputs, save to files in
    /tmp/
    and summarise in-chat.
  • 始终使用内嵌markdown链接引用网页来源的事实
    [来源标题](https://...)
  • 只要使用了搜索/提取/研究的输出,结尾必须附上来源列表
  • 优先使用官方/原始来源(如果可用)。
  • 对于长输出,保存到
    /tmp/
    目录下的文件中,并在对话中提供摘要。

Search (default web lookup)

搜索(默认网页查询)

Use Search for fast, cost-effective answers with citations.
使用搜索功能获取带引用的快速、高性价比答案。

Command template

命令模板

bash
parallel-cli search "$OBJECTIVE"   --mode agentic   --max-results 10   --json
Add any of these only when relevant:
  • --after-date YYYY-MM-DD
    (freshness constraint)
  • --include-domains a.com b.org
    (restrict sources)
  • --exclude-domains spam.com
    (block sources)
  • one or more
    -q "keyword query"
    flags (extra keyword probes)
  • -o "/tmp/$SLUG.search.json"
    (save full JSON to a file)
bash
parallel-cli search "$OBJECTIVE"   --mode agentic   --max-results 10   --json
仅在相关时添加以下参数:
  • --after-date YYYY-MM-DD
    (新鲜度限制)
  • --include-domains a.com b.org
    (限制来源域名)
  • --exclude-domains spam.com
    (屏蔽来源域名)
  • 一个或多个
    -q "keyword query"
    标志(额外关键词查询)
  • -o "/tmp/$SLUG.search.json"
    (将完整JSON保存到文件)

Parse + respond

解析与回复

From the JSON results, extract title, url, and any publish_date / excerpt fields. Answer the user’s question, and cite each claim inline.
从JSON结果中提取标题url以及任何发布日期/摘要字段。 回答用户的问题,并为每个主张添加内嵌引用。

Extract (read one or more URLs)

提取(读取一个或多个URL)

Use Extract when you need the actual contents of specific URLs (webpages, PDFs, JS-heavy sites).
当你需要获取特定URL(网页、PDF、JS渲染网站)的实际内容时,使用提取功能。

Command template

命令模板

bash
parallel-cli extract "$URL" --json
Add when relevant:
  • --objective "Focus area"
    (e.g., pricing, API usage, constraints)
  • --full-content
    (only if the user needs the whole page)
  • --no-excerpts
    (if you only want full content)
  • -o "/tmp/$SLUG.extract.json"
    (save full JSON to a file)
bash
parallel-cli extract "$URL" --json
相关时添加以下参数:
  • --objective "Focus area"
    (重点领域,例如定价、API使用、限制条件)
  • --full-content
    (仅当用户需要完整页面内容时使用)
  • --no-excerpts
    (如果只需要完整内容)
  • -o "/tmp/$SLUG.extract.json"
    (将完整JSON保存到文件)

Respond

回复

  • If the user asked for a summary, summarise with citations to the extracted URL.
  • If the user asked for the verbatim text, provide the extracted markdown only if it is reasonably sized; otherwise provide the key sections + offer to read more from the saved output.
  • 如果用户要求摘要,则生成摘要并引用提取的URL。
  • 如果用户要求原文文本,仅在内容长度合理时提供提取的markdown;否则提供关键章节内容,并告知可从保存的输出中读取更多内容。

Deep research (only when explicitly requested)

深度研究(仅在明确请求时使用)

Deep research is slower and may cost more than Search. Use it only when the user explicitly wants depth.
深度研究比搜索更慢,成本可能更高。仅在用户明确要求深度内容时使用。

Step 1 — start (always async)

步骤1 — 启动(始终异步)

bash
parallel-cli research run "$QUESTION" --processor pro-fast --no-wait --json
Parse
run_id
(and any monitoring URL) from JSON and tell the user the run started.
bash
parallel-cli research run "$QUESTION" --processor pro-fast --no-wait --json
从JSON中解析
run_id
(以及任何监控URL),并告知用户任务已启动。

Step 2 — poll (bounded timeout)

步骤2 — 轮询(有界超时)

Choose a short slug filename (lowercase-hyphen), then:
bash
parallel-cli research poll "$RUN_ID" -o "/tmp/$SLUG" --timeout 540
  • Share the executive summary printed by the poll command.
  • Mention the output files:
    • /tmp/$SLUG.md
    • /tmp/$SLUG.json
If polling times out, re-run the same poll command — the run continues server-side.
选择一个简短的小写连字符命名的文件名,然后执行:
bash
parallel-cli research poll "$RUN_ID" -o "/tmp/$SLUG" --timeout 540
  • 分享轮询命令输出的执行摘要
  • 提及输出文件:
    • /tmp/$SLUG.md
    • /tmp/$SLUG.json
如果轮询超时,重新执行相同的轮询命令——任务会在服务器端继续运行。

Enrich (CSV/JSON or inline data)

增强(CSV/JSON或内联数据)

Use Enrich to add web-sourced columns to structured data.
使用增强功能为结构化数据添加网页来源的列。

Step 1 — (optional) suggest columns

步骤1 —(可选)建议列

bash
parallel-cli enrich suggest "$INTENT" --json
Use this when the user knows the goal but not the exact output schema.
bash
parallel-cli enrich suggest "$INTENT" --json
当用户知道目标但不确定确切输出模式时使用此命令。

Step 2 — run (always async for large jobs)

步骤2 — 运行(大型任务始终异步)

For CSV:
bash
parallel-cli enrich run   --source-type csv   --source "input.csv"   --target "/tmp/enriched.csv"   --source-columns '[{"name":"company","description":"Company name"}]'   --intent "$INTENT"   --no-wait --json
For inline JSON rows:
bash
parallel-cli enrich run   --data '[{"company":"Google"},{"company":"Apple"}]'   --target "/tmp/enriched.csv"   --intent "$INTENT"   --no-wait --json
Parse
taskgroup_id
from JSON.
针对CSV:
bash
parallel-cli enrich run   --source-type csv   --source "input.csv"   --target "/tmp/enriched.csv"   --source-columns '[{"name":"company","description":"Company name"}]'   --intent "$INTENT"   --no-wait --json
针对内联JSON行:
bash
parallel-cli enrich run   --data '[{"company":"Google"},{"company":"Apple"}]'   --target "/tmp/enriched.csv"   --intent "$INTENT"   --no-wait --json
从JSON中解析
taskgroup_id

Step 3 — poll

步骤3 — 轮询

bash
parallel-cli enrich poll "$TASKGROUP_ID" --timeout 540 --json
After completion:
  • Tell the user the output file path (the
    --target
    you chose).
  • Preview a few rows (using file read tools if available) and report row counts.
If poll times out, re-run it — the job continues server-side.
bash
parallel-cli enrich poll "$TASKGROUP_ID" --timeout 540 --json
完成后:
  • 告知用户输出文件路径(你选择的
    --target
    路径)。
  • 预览几行内容(如果可用,使用文件读取工具)并报告行数。
如果轮询超时,重新执行——任务会在服务器端继续运行。

FindAll (entity discovery)

FindAll(实体发现)

Use FindAll when the user wants you to discover a set of entities (e.g., “AI startups in healthcare”, “roofing companies in Charlotte”, “YC devtools companies”).
当用户希望你发现一组实体时(例如“医疗领域的AI初创公司”、“夏洛特的屋顶公司”、“YC旗下的开发工具公司”),使用FindAll功能。

Step 1 — run

步骤1 — 运行

bash
parallel-cli findall run "$OBJECTIVE" --generator core --match-limit 25 --no-wait --json
Useful options:
  • --dry-run --json
    to preview schema before spending money
  • --exclude '[{"name":"Example Corp","url":"example.com"}]'
    to avoid known entities
  • --generator preview|base|core|pro
    (core default; pro for hardest queries)
Parse
run_id
from JSON.
bash
parallel-cli findall run "$OBJECTIVE" --generator core --match-limit 25 --no-wait --json
实用选项:
  • --dry-run --json
    (在消耗成本前预览模式)
  • --exclude '[{"name":"Example Corp","url":"example.com"}]'
    (排除已知实体)
  • --generator preview|base|core|pro
    (默认core;难度最高的查询使用pro)
从JSON中解析
run_id

Step 2 — poll + fetch results

步骤2 — 轮询并获取结果

bash
parallel-cli findall poll "$RUN_ID" --json
parallel-cli findall result "$RUN_ID" --json
Respond with:
  • total entities found
  • a clean list/table of the best matches (name + URL + key attributes)
  • any caveats about ambiguous matches
bash
parallel-cli findall poll "$RUN_ID" --json
parallel-cli findall result "$RUN_ID" --json
回复内容包括:
  • 发现的实体总数
  • 匹配度最高的实体的整洁列表/表格(名称 + URL + 关键属性)
  • 关于模糊匹配的注意事项

Monitor (web change tracking)

监控(网页变化跟踪)

Use Monitor when the user wants ongoing tracking.
Create:
bash
parallel-cli monitor create "$OBJECTIVE" --cadence daily --json
Optional:
  • --cadence hourly|daily|weekly|every_two_weeks
  • --webhook https://example.com/hook
    (deliver events externally)
  • --output-schema '<JSON schema string>'
    (structured events)
Manage:
bash
parallel-cli monitor list --json
parallel-cli monitor get "$MONITOR_ID" --json
parallel-cli monitor update "$MONITOR_ID" --cadence weekly --json
parallel-cli monitor delete "$MONITOR_ID"
parallel-cli monitor events "$MONITOR_ID" --json
parallel-cli monitor simulate "$MONITOR_ID" --json
Respond with the monitor id and how to retrieve events (or confirm webhook delivery).
当用户需要持续跟踪时,使用监控功能。
创建监控任务:
bash
parallel-cli monitor create "$OBJECTIVE" --cadence daily --json
可选参数:
  • --cadence hourly|daily|weekly|every_two_weeks
  • --webhook https://example.com/hook
    (向外部推送事件)
  • --output-schema '<JSON schema string>'
    (结构化事件)
管理监控任务:
bash
parallel-cli monitor list --json
parallel-cli monitor get "$MONITOR_ID" --json
parallel-cli monitor update "$MONITOR_ID" --cadence weekly --json
parallel-cli monitor delete "$MONITOR_ID"
parallel-cli monitor events "$MONITOR_ID" --json
parallel-cli monitor simulate "$MONITOR_ID" --json
回复监控ID以及获取事件的方式(或确认webhook推送状态)。

Reference material

参考资料

  • Copy/paste command templates and patterns:
    references/command-templates.md
  • Troubleshooting common failures:
    references/troubleshooting.md
  • 复制粘贴命令模板与模式:
    references/command-templates.md
  • 常见故障排查:
    references/troubleshooting.md