Perplexity API in Make.com: Complete Tutorial with Working Examples (2026)

Perplexity is the API I reach for whenever a Make.com scenario needs live web research as part of its flow. It’s not a general-purpose LLM like Claude or Gemini — it’s specifically built to search the web, read the results, and return a grounded answer with citations.

This guide covers how to connect the Perplexity API to Make.com, pick the right model, handle the output, and build two real workflows I use.

What Perplexity Actually Does

Perplexity’s API runs what they call Sonar models. Each API call does three things internally:

1. Searches the live web for relevant sources
2. Reads and synthesises the top results
3. Returns an answer with source citations

You don’t write your own search queries — you ask a question in natural language, and Perplexity handles the search strategy. The trade-off: less control than rolling your own search, but way less engineering.

For automation use cases — daily market summaries, competitor monitoring, news briefings, live pricing research — this is exactly the right tradeoff.

Setting Up the Connection

Step 1: Get a Perplexity API Key

Go to perplexity.ai/settings/api. You’ll need to sign up if you haven’t already, then add at least $5 of credit — Perplexity is pay-as-you-go, no free tier.

Click Generate → copy the key → store it somewhere safe. It starts with pplx-.

Step 2: Connect to Make.com

Make.com doesn’t (as of April 2026) have a dedicated Perplexity module in the main directory. Instead, you use the generic HTTP module to call Perplexity’s API directly. It’s five minutes of work.

1. In your scenario, add an HTTP module → Make a request
2. Configure it as: - URL: https://api.perplexity.ai/chat/completions - Method: POST - Headers:
   • Authorization: Bearer YOUR_PERPLEXITY_KEY
   • Content-Type: application/json
   • Body type: Raw → JSON
   • Request content:

{
  "model": "sonar",
  "messages": [
    {"role": "user", "content": "Your question here"}
  ]
}

Click Run once and you should get a response containing the answer, sources, and token usage.

Which Model to Use

As of April 2026, Perplexity offers several Sonar variants:

• sonar — fast, basic web search grounded answers
• sonar-pro — deeper research, longer context
• sonar-reasoning — adds chain-of-thought reasoning before answering

Pricing (per million tokens, roughly as of April 2026):

• sonar: ~$1 input / ~$1 output
• sonar-pro: ~$3 input / ~$15 output
• sonar-reasoning: ~$1 input / ~$5 output

For most automation work — “what’s the latest news about X” — sonar is enough. Reach for sonar-pro only when you need depth (long research reports, comprehensive competitor analysis).

(Always verify current pricing at docs.perplexity.ai — changes regularly.)

Parsing the Response

A Perplexity response looks like this:

{
  "id": "req_...",
  "model": "sonar",
  "choices": [
    {
      "message": {
        "content": "The answer text here...",
        "role": "assistant"
      }
    }
  ],
  "citations": [
    "https://source1.com/article",
    "https://source2.com/report"
  ],
  "usage": {
    "prompt_tokens": 42,
    "completion_tokens": 180,
    "total_tokens": 222
  }
}

In Make.com, after the HTTP module runs, you map:

• Answer text: choices[].message.content (array index 1)
• Sources: citations[] — this is an array you can iterate over
• Tokens used: usage.total_tokens (useful for cost monitoring)

Example 1: Daily Industry News Brief

This is a scenario I run at 8am every weekday. Delivers a 5-bullet summary of my industry’s news to my email.

Modules:

1. Schedule trigger — runs at 8:00am weekdays
2. HTTP → Perplexity — prompt:

   “Give me the top 5 news stories from the last 24 hours about Australian premium chauffeur and transport services. For each, write one sentence summarising what happened and why it matters.”

3. Email module — sends the content to my inbox with subject “Industry brief — [today’s date]”

Total runtime: about 8 seconds. Cost: roughly $0.002 per run (sonar model). Monthly cost: under 10 cents.

Yes — under 10 cents for a personalised daily news brief. That’s why I love the Perplexity API.

Example 2: Competitor Price Monitoring

This one watches a competitor’s public pricing page and alerts me if anything changes.

Modules:

1. Schedule trigger — daily
2. HTTP → Perplexity (sonar-pro) — prompt:

   “Visit [competitor-url] and return their current pricing in JSON format with route and price fields. Return ONLY the JSON.”

3. JSON → Parse the response
4. Data Store lookup — retrieves last known prices
5. Filter — only continue if any price changed
6. Telegram / Email — alert with old vs new prices

The key pattern: Perplexity for live fetching + grounding, Data Store for state. Works better than scraping because Perplexity handles HTML parsing internally.

When Perplexity Is the Right Tool

• Time-sensitive information — news, prices, events, anything where your data freshness matters
• Research that needs citations — reports where you need to show sources
• Live web content — pulling from multiple pages without building your own scraper

When Perplexity Is the Wrong Tool

• Deterministic processing — if you have structured data and need to transform it, use Claude or Gemini instead
• Long-form creative writing — Perplexity’s outputs are research-flavoured; for creative content, use Claude
• Complex multi-step reasoning — Perplexity answers directly; if you need “think step by step”, use Claude Opus

Cost Optimisation Tips

1. Use sonar, not sonar-pro, by default. Most queries don’t need the deeper variant.
2. Cache when possible. If you’re asking the same question multiple times a day, store the answer in a Make.com Data Store and refresh only when needed.
3. Keep prompts specific. “Give me 5 bullets” is cheaper and faster than “tell me everything about X.”
4. Set hard spend limits in your Perplexity dashboard.

Next Steps

Perplexity + Claude together is the “API Trinity” backbone for research-driven automation. The Implementation Blueprint ($29) includes full Make.com scenarios using this pattern — daily briefs, competitor monitoring, meeting prep from calendar invites, all with cost-optimised model routing.

The free Quick Start uses Claude only — it’s a cleaner first scenario. Once you’ve got that running, Perplexity is the natural next addition.

---

Last updated: 20 April 2026.