parallel-web
Skill80 starsChanged 8 months agoMITView on GitHub
Search the web, extract URL content, and run deep research using the Parallel Chat API and Extract API. Use for ALL web searches, research queries, and general information gathering. Provides synthesized summaries with citations.
- Reads credentials
- Installs packages
7 other projects use a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- K-Dense-AI/claude-scientific-writertheir own version, first seen Nov 7, 2026
- MarieLynneBlock/arcanum-artifexedited, first seen Nov 1, 2026
- yigityildiz0/evidence-research-skillstheir own version, first seen Oct 25, 2026
- patoangolano/bio-stacktheir own version, first seen Oct 15, 2026
- Luooops/skillstheir own version, first seen Oct 13, 2026
- uppifyagency/boilerplate-saas-2026their own version, first seen Jun 2, 2026
Other versions7 projects use a version of this.
Word for word · 1
- LeonChaoX/qinyan-academic-skills★ 910100% same
Edited · 1
- MarieLynneBlock/arcanum-artifex★ 498% same
Wrote their own version · 5
- K-Dense-AI/claude-scientific-writer★ 2.6ksame name
- Luooops/skills★ 1same name
- uppifyagency/boilerplate-saas-2026★ 1same name
- patoangolano/bio-stack★ 0same name
- yigityildiz0/evidence-research-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository180 other files it gives its agents.
- bgpt-paper-searchSkill
- biorxiv-databaseSkill
- citation-managementSkill
- literature-reviewSkill
- openalex-databaseSkill
- perplexity-searchSkill
- pubmed-databaseSkill
- pyzoteroSkill
- research-lookupSkill
- medical-imaging-reviewSkill
- paper-2-webSkill
- peer-reviewSkill
- research-proposalSkill
- scientific-writingSkill
- venue-templatesSkill
- generate-imageSkill
- infographicsSkill
- latex-postersSkill
- markdown-mermaid-writingSkill
- paper-slide-deckSkill
- pptx-postersSkill
- scientific-schematicsSkill
- scientific-slidesSkill
- scientific-visualizationSkill
- consciousness-councilSkill
- dhdna-profilerSkill
- hypothesis-generationSkill
- research-grantsSkill
- scholar-evaluationSkill
- scientific-brainstormingSkill
- scientific-critical-thinkingSkill
- what-if-oracleSkill
- anndataSkill
- arboretoSkill
- biopythonSkill
- bioservicesSkill
- cellxgene-censusSkill
- deeptoolsSkill
- etetoolkitSkill
- flowioSkill
- genimlSkill
- ggetSkill
- gtarsSkill
- lamindbSkill
- phylogeneticsSkill
- pydeseq2Skill
- pysamSkill
- scanpySkill
- scikit-bioSkill
- scveloSkill
- scvi-toolsSkill
- tiledbvcfSkill
- zarr-pythonSkill
- datamolSkill
- deepchemSkill
- diffdockSkill
- matchmsSkill
- medchemSkill
- molecular-dynamicsSkill
- molfeatSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 3 are distinctive, and 5 are very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 26
Use this skill for ALL of the following:
- Very specific: line 50
Search the web via the Parallel Chat API (base model) and get a synthesized summary with cited sources.
- Very specific: line 78
Run comprehensive multi-source research via the Parallel Chat API (core model) that produces detailed intelligence reports with citations.
- Distinctive: line 224
Reproducibility: Every claim in the final document can be traced back to its raw source material
- Very specific: line 225
Context Window Recovery: If context is compacted mid-task, saved results can be re-read from sources/
- Very specific: line 228
Cost Efficiency: Avoid redundant API calls by checking sources/ for existing results
- Distinctive: line 233
When saving research results, always log:
- Very specific: line 242
Before calling parallel_web.py, check if a relevant result already exists in sources/:
“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Parallel Web Systems API
Overview
This skill provides access to Parallel Web Systems APIs for web search, deep research, and content extraction. It is the primary tool for all web-related operations in the scientific writer workflow.
Primary interface: Parallel Chat API (OpenAI-compatible) for search and research. Secondary interface: Extract API for URL verification and special cases only.
API Documentation: https://docs.parallel.ai
API Key: https://platform.parallel.ai
Environment Variable: PARALLEL_API_KEY
When to Use This Skill
Use this skill for ALL of the following:
- Web Search: Any query that requires searching the internet for information
- Deep Research: Comprehensive research reports on any topic
- Market Research: Industry analysis, competitive intelligence, market data
- Current Events: News, recent developments, announcements
- Technical Information: Documentation, specifications, product details
- Statistical Data: Market sizes, growth rates, industry figures
- General Information: Company profiles, facts, comparisons
Use Extract API only for:
- Citation verification (confirming a specific URL's content)
- Special cases where you need raw content from a known URL
Do NOT use this skill for:
- Academic-specific paper searches (use
research-lookupwhich routes to Perplexity for purely academic queries) - Google Scholar / PubMed database searches (use
citation-managementskill)
Two Capabilities
1. Web Search (search command)
Search the web via the Parallel Chat API (base model) and get a synthesized summary with cited sources.
Best for: General web searches, current events, fact-finding, technical lookups, news, market data.
# Basic search
python scripts/parallel_web.py search "latest advances in quantum computing 2025"
# Use core model for more complex queries
python scripts/parallel_web.py search "compare EV battery chemistries NMC vs LFP" --model core
# Save results to file
python scripts/parallel_web.py search "renewable energy policy updates" -o results.txt
# JSON output for programmatic use
python scripts/parallel_web.py search "AI regulation landscape" --json -o results.jsonKey Parameters:
objective: Natural language description of what you want to find--model: Chat model to use (basedefault, orcorefor deeper research)-o: Output file path--json: Output as JSON
Response includes: Synthesized summary organized by themes, with inline citations and a sources list.
2. Deep Research (research command)
Run comprehensive multi-source research via the Parallel Chat API (core model) that produces detailed intelligence reports with citations.
Best for: Market research, comprehensive analysis, competitive intelligence, technology surveys, industry reports, any research question requiring synthesis of multiple sources.
# Default deep research (core model)
python scripts/parallel_web.py research "comprehensive analysis of the global EV battery market"
# Save research report to file
python scripts/parallel_web.py research "AI adoption in healthcare 2025" -o report.md
# Use base model for faster, lighter research
python scripts/parallel_web.py research "latest funding rounds in AI startups" --model base
# JSON output
python scripts/parallel_web.py research "renewable energy storage market in Europe" --json -o data.jsonKey Parameters:
query: Research question or topic--model: Chat model to use (coredefault for deep research, orbasefor faster results)-o: Output file path--json: Output as JSON
3. URL Extraction (extract command) — Verification Only
Extract content from specific URLs. Use only for citation verification and special cases.
For general research, use search or research instead.
# Verify a citation's content
python scripts/parallel_web.py extract "https://example.com/article" --objective "key findings"
# Get full page content for verification
python scripts/parallel_web.py extract "https://docs.example.com/api" --full-content
# Save extraction to file
python scripts/parallel_web.py extract "https://paper-url.com" --objective "methodology" -o extracted.mdModel Selection Guide
The Chat API supports two research models. Use base for most searches and core for deep research.
| Model | Latency | Strengths | Use When |
base | 15s-100s | Standard research, factual queries | Web searches, quick lookups |
core | 60s-5min | Complex research, multi-source synthesis | Deep research, comprehensive reports |
Recommendations:
searchcommand defaults tobase— fast, good for most queriesresearchcommand defaults tocore— thorough, good for comprehensive reports- Override with
--modelwhen you need different depth/speed tradeoffs
Python API Usage
Search
from parallel_web import ParallelSearch
searcher = ParallelSearch()
result = searcher.search(
objective="Find latest information about transformer architectures in NLP",
model="base",
)
if result["success"]:
print(result["response"]) # Synthesized summary
for src in result["sources"]:
print(f" {src['title']}: {src['url']}")Deep Research
from parallel_web import ParallelDeepResearch
researcher = ParallelDeepResearch()
result = researcher.research(
query="Comprehensive analysis of AI regulation in the EU and US",
model="core",
)
if result["success"]:
print(result["response"]) # Full research report
print(f"Citations: {result['citation_count']}")Extract (Verification Only)
from parallel_web import ParallelExtract
extractor = ParallelExtract()
result = extractor.extract(
urls=["https://docs.example.com/api-reference"],
objective="API authentication methods and rate limits",
)
if result["success"]:
for r in result["results"]:
print(r["excerpts"])MANDATORY: Save All Results to Sources Folder
Every web search and deep research result MUST be saved to the project's sources/ folder.
This ensures all research is preserved for reproducibility, auditability, and context window recovery.
Saving Rules
| Operation | -o Flag Target | Filename Pattern |
| Web Search | sources/search_<topic>.md | search_YYYYMMDD_HHMMSS_<brief_topic>.md |
| Deep Research | sources/research_<topic>.md | research_YYYYMMDD_HHMMSS_<brief_topic>.md |
| URL Extract | sources/extract_<source>.md | extract_YYYYMMDD_HHMMSS_<brief_source>.md |
How to Save (Always Use -o Flag)
CRITICAL: Every call to parallel_web.py MUST include the -o flag pointing to the sources/ folder.
# Web search — ALWAYS save to sources/
python scripts/parallel_web.py search "latest advances in quantum computing 2025" \
-o sources/search_20250217_143000_quantum_computing.md
# Deep research — ALWAYS save to sources/
python scripts/parallel_web.py research "comprehensive analysis of the global EV battery market" \
-o sources/research_20250217_144000_ev_battery_market.md
# URL extraction (verification only) — save to sources/
python scripts/parallel_web.py extract "https://example.com/article" --objective "key findings" \
-o sources/extract_20250217_143500_example_article.mdWhy Save Everything
- Reproducibility: Every claim in the final document can be traced back to its raw source material
- Context Window Recovery: If context is compacted mid-task, saved results can be re-read from
sources/ - Audit Trail: The
sources/folder provides complete transparency into how information was gathered - Reuse Across Sections: Saved research can be referenced by multiple sections without duplicate API calls
- Cost Efficiency: Avoid redundant API calls by checking
sources/for existing results - Peer Review Support: Reviewers can verify the research backing every claim
Logging
When saving research results, always log:
[HH:MM:SS] SAVED: Search results to sources/search_20250217_143000_quantum_computing.md
[HH:MM:SS] SAVED: Deep research report to sources/research_20250217_144000_ev_battery_market.mdBefore Making a New Query, Check Sources First
Before calling parallel_web.py, check if a relevant result already exists in sources/:
ls sources/ # Check existing saved resultsIntegration with Scientific Writer
Routing Table
| Task | Tool | Command |
| Web search (any) | parallel_web.py search | python scripts/parallel_web.py search "query" -o sources/search_<topic>.md |
| Deep research | parallel_web.py research | python scripts/parallel_web.py research "query" -o sources/research_<topic>.md |
| Citation verification | parallel_web.py extract | python scripts/parallel_web.py extract "url" -o sources/extract_<source>.md |
| Academic paper search | research_lookup.py | Routes to Perplexity sonar-pro-search |
| DOI/metadata lookup | parallel_web.py extract | Extract from DOI URLs (verification) |
When Writing Scientific Documents
- Before writing any section, use
searchorresearchto gather background information — save results tosources/ - For academic citations, use
research-lookup(which routes academic queries to Perplexity) — save results tosources/ - For citation verification (confirming a specific URL), use
parallel_web.py extract— save results tosources/ - For current market/industry data, use
parallel_web.py research --model core— save results tosources/ - Before any new query, check
sources/for existing results to avoid duplicate API calls
Environment Setup
# Required: Set your Parallel API key
export PARALLEL_API_KEY="your_api_key_here"
# Required Python packages
pip install openai # For Chat API (search/research)
pip install parallel-web # For Extract API (verification only)Get your API key at https://platform.parallel.ai
Error Handling
The script handles errors gracefully and returns structured error responses:
{
"success": false,
"error": "Error description",
"timestamp": "2025-02-14 12:00:00"
}Common issues:
PARALLEL_API_KEY not set: Set the environment variableopenai not installed: Runpip install openaiparallel-web not installed: Runpip install parallel-web(only needed for extract)Rate limit exceeded: Wait and retry (default: 300 req/min for Chat API)
Complementary Skills
| Skill | Use For |
research-lookup | Academic paper searches (routes to Perplexity for scholarly queries) |
citation-management | Google Scholar, PubMed, CrossRef database searches |
literature-review | Systematic literature reviews across academic databases |
scientific-schematics | Generate diagrams from research findings |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

