name	perplexity-research
description	Real-time web research via Perplexity Sonar API. Use when user needs live web data, current documentation, market research, competitor analysis, or asks 'what's the latest on'.
allowed-tools	["Bash","Read"]

Perplexity Research Skill

Overview

Calls the Perplexity Sonar API for real-time web research, returning cited answers that feed into Claude/Amp workflows.

Prerequisites

PERPLEXITY_API_KEY set in .env (already configured ✅)
Python 3 with requests package (already installed ✅)

Usage

From CLI

# Quick research query
python3 scripts/perplexity_orchestrator.py --query "your question here"

# Use stronger model
python3 scripts/perplexity_orchestrator.py --query "complex topic" --model sonar-pro

# Raw JSON output for piping
python3 scripts/perplexity_orchestrator.py --query "topic" --json

# Self-test
python3 scripts/perplexity_orchestrator.py --test

From Claude/Amp

When a user asks for live web research, run:

source .env && export PERPLEXITY_API_KEY && python3 scripts/perplexity_orchestrator.py --query "<user query>"

Available Models

Model	Use Case
`sonar`	Fast, general research (default)
`sonar-pro`	Deeper analysis, longer answers
`sonar-reasoning`	Step-by-step reasoning with citations
`sonar-reasoning-pro`	Complex multi-step research

Integration with Claude/Amp Skills

This skill augments existing workflows:

Growth pipeline: Feed market research into growth_content_pipeline.py
Release context: Research App Store guideline changes before releases
North Star tracking: Research competitor benchmarks and industry baselines

Limitations

Perplexity Computer (cloud worker) is NOT controllable via this API — it's a separate Perplexity product requiring their web UI
Comet browser is a Perplexity-internal component, not externally launchable
API rate limits apply per your Perplexity subscription tier
HTTP 401 from api.perplexity.ai may mean insufficient_quota (plan/billing), not a missing key — check Perplexity API settings. Scripts must load .env via repo_dotenv (or export PERPLEXITY_API_KEY) because bare os.environ in CI/agents often has no key.

Perplexity Research Skill

Overview

Calls the Perplexity Sonar API for real-time web research, returning cited answers that feed into Claude/Amp workflows.

Prerequisites

PERPLEXITY_API_KEY set in .env (already configured ✅)

Python 3 with requests package (already installed ✅)

Usage

From CLI

# Quick research query python3 scripts/perplexity_orchestrator.py --query "your question here" # Use stronger model python3 scripts/perplexity_orchestrator.py --query "complex topic" --model sonar-pro # Raw JSON output for piping python3 scripts/perplexity_orchestrator.py --query "topic" --json # Self-test python3 scripts/perplexity_orchestrator.py --test

From Claude/Amp

When a user asks for live web research, run:

source .env && export PERPLEXITY_API_KEY && python3 scripts/perplexity_orchestrator.py --query "<user query>"

Available Models

Model

Use Case

sonar

Fast, general research (default)

sonar-pro

Deeper analysis, longer answers

sonar-reasoning

Step-by-step reasoning with citations

sonar-reasoning-pro

Complex multi-step research

Integration with Claude/Amp Skills

This skill augments existing workflows:

Growth pipeline: Feed market research into growth_content_pipeline.py

Release context: Research App Store guideline changes before releases

North Star tracking: Research competitor benchmarks and industry baselines

Limitations

Perplexity Computer (cloud worker) is NOT controllable via this API — it's a separate Perplexity product requiring their web UI

Comet browser is a Perplexity-internal component, not externally launchable

API rate limits apply per your Perplexity subscription tier

HTTP 401 from api.perplexity.ai may mean insufficient_quota (plan/billing), not a missing key — check Perplexity API settings. Scripts must load .env via repo_dotenv (or export PERPLEXITY_API_KEY) because bare os.environ in CI/agents often has no key.

perplexity-research