Perplexity
Search-GroundedThe Sonar API — purpose-built for search-grounded retrieval and cited answers.
Perplexity has graduated from a search interface into a four-tier model provider with the Sonar API. Each tier is tuned for retrieval-augmented generation, returning answers with inline source citations that eliminate hallucinations on factual queries. The Pro plan ($20/month) is the practical sweet spot for individual researchers; Perplexity Max ($200/month) provides an 'agentic computer' that orchestrates 19 different models behind the scenes.
Models
Sonar
High-volume cited chat — $5–12 per 1K requests
Context Window
128K
Input Price
$1.00 / 1M
Output Price
$1.00 / 1M
Input Types
Text
Output Types
Text with citations
Sonar Pro
RecommendedProduction answer engines — $6–14 per 1K requests
Context Window
200K
Input Price
$3.00 / 1M
Output Price
$15.00 / 1M
Input Types
Text
Output Types
Text with citations
Sonar Reasoning
Complex search logic — chain-of-thought + retrieval
Context Window
128K
Input Price
$1.00 / 1M
Output Price
$5.00 / 1M
Input Types
Text
Output Types
Text with citations
Sonar Deep Research
NewLong-form report generation — $5 per searches
Context Window
128K
Input Price
$2.00 / 1M
Output Price
$8.00 / 1M
Input Types
Text
Output Types
Text with citations
Use Cases
Competitive Intelligence
Continuously monitor competitors and regulatory shifts with live web citations — auditable for compliance.
Customer-Facing Q&A Agents
Sonar Pro powers production answer engines that ground every response in real-time product info, reviews, and specs.
Long-Form Research Reports
Sonar Deep Research executes multi-step autonomous research and produces structured reports from a single brief.
Compliance & Regulatory
Retrieve and cite current regulations, court rulings, and policy documents — every claim is sourced.
Why use Perplexity on Nagent?
Nagent adds enterprise orchestration, observability, and workflow automation on top of Perplexity's raw model capabilities.
No need to build a custom web-search tool — Sonar handles retrieval natively
Cited sources in every response — auditable, hallucination-resistant outputs
Combine Sonar with Nagent's CRM and notification tools in one workflow
Deep Research model handles autonomous multi-step research without orchestration code
How to access Perplexity on Nagent
Open Agent Studio
Navigate to Agent Studio in your Nagent workspace.
Select Perplexity
Choose the Sonar tier that matches your workload — Pro for production, Reasoning for complex queries, Deep Research for reports.
Build Your Research Agent
Define query templates and output format — Sonar handles retrieval and citation automatically.
Common questions about Perplexity
Real buyer and developer questions, answered. Click any item to expand.
Sonar vs Sonar Pro — when do I need Pro?
Sonar is right for high-volume cited chat. Sonar Pro is right for production answer engines where retrieval depth and reasoning quality both matter. Pro costs roughly 3x more on output tokens but the answer quality jump is significant.
Does Sonar Deep Research replace a research analyst?
For initial drafting and source aggregation — largely, yes. One call produces what a junior analyst would deliver in 5–15 minutes. For final synthesis and judgement calls, pair the output with a human reviewer.
How are citations stored in my Nagent agent's outputs?
Citations come back as structured metadata on the response. Nagent's KARMIC ledger persists them so you can audit any factual claim back to its source URL — useful for compliance reviews and regulated reporting.
Can I use Perplexity inside a Claude or GPT-5 agent workflow?
Yes — many agents use Perplexity as a tool call from a Claude or GPT brain. The LLM decides when to call Sonar for grounded retrieval, then synthesises the answer using its own reasoning. Best of both worlds.
Ready to use Perplexity inside your agents?
Get started in minutes — no API key management required.