Evaluation and observability platform for LLM apps, built on the DeepEval framework, with tracing, datasets and regression tests.

Perplexity Sonar API
Pay-as-you-go Perplexity API for web-grounded answers, ranked search results, embeddings and typed decisions, with an OpenAI-compatible endpoint.
What is Perplexity Sonar API?
The Perplexity Sonar API is the developer platform behind Perplexity's web research abilities. Sonar chat completions return web-grounded answers through an OpenAI-compatible interface, so existing OpenAI client libraries work by pointing them at Perplexity's base URL, and native Python and TypeScript SDKs are also published. The documentation notes that Sonar chat completions are being succeeded by the Agent API, which returns answers with built-in citations and can call third-party models from OpenAI, Anthropic, Google, xAI and others with web search tools. Other APIs include a Search API for ranked web results with filters, an Embeddings API for semantic search and retrieval-augmented generation, a Decisions API for yes/no or multiple-choice answers with probabilities, and an Orchestrator API for hosted open-weight models. Pricing is pay as you go with no subscription: for example, the Search API is $5.00 per 1K requests, and Agent API web search is $0.0025 per invocation, plus model tokens. Developers building research assistants, RAG pipelines or agents use it.
Use cases
- Add web-grounded answers to an app
- Fetch ranked web search results
- Build RAG with hosted embeddings
- Classify text or images with probabilities
- Build agents with web search tools
- Route calls to multiple model providers
Who it's for
Why use Perplexity Sonar API
- Reuses existing OpenAI client code, so a first call takes minutes rather than a migration project.
- Ties answers to web sources with citations, which helps apps show where a claim came from.
- Charges only for what you use, with no subscription or minimum shown in the pricing docs.
- Combines search, embeddings and decision-style classification under one account and key.
- Lets you try several model providers through one Agent API without separate vendor contracts.
Perplexity Sonar API features
- Sonar chat completions with web-grounded answers and citations
- Agent API with web search, URL fetch and a sandbox tool
- Third-party models from OpenAI, Anthropic, Google, xAI and others
- Search API with filtering and a Fast Search option
- Embeddings API with standard and contextualized models
- Decisions API returning typed answers with probabilities
- Orchestrator API for hosted open-weight models
- OpenAI-compatible endpoint
- Python and TypeScript SDKs
- Credits purchasable through AWS Marketplace
Perplexity Sonar API integrations
How to get started with Perplexity Sonar API
- Generate an API key on the API Keys page of the Perplexity API Console.
- Install the Perplexity SDK or an OpenAI client and set the PERPLEXITY_API_KEY environment variable.
- Send a request to the Agent API, Sonar endpoint, Search API or another endpoint.
- Read the response, including citations and exact token counts returned with each call.
- Buy credits or set up auto top-off, then watch usage per key.
Perplexity Sonar API limitations
- Costs scale with usage, combining model tokens and per-call tool fees
- Sonar chat completions are being replaced by the Agent API, and the docs give a support end date of September 27, 2026
- Anthropic models require a max_output_tokens parameter or the API returns an error
- Some Agent API model rates change above certain context lengths
Perplexity Sonar API pricing
Perplexity Sonar API has no free plan.
Perplexity Sonar API plans
Search API
$5.00 per 1K requests
- Raw web search results
- Advanced filtering
- No token-based charges
Search API with Fast Search
$1.00 per 1K requests
- Lower-latency web search results
- Set search_type to fast
Agent API web search
$0.0025 per invocation
plus model token costs
- Standard web search tool
- Fast Search variant at $0.001 per invocation
Embeddings API
$0.004 per 1M tokens
pplx-embed-v1-0.6b
- Standard and contextualized models
- Up to 2560 dimensions
Decisions API
$0.02 per 1M input tokens
pplx-decider-v1.2
- Yes/no, multiple-choice and scored answers
- Output tokens are free
- No per-request fee
Orchestrator API
Price on vendor site
per token at each model's rate
- Open-weight models hosted by Perplexity
- No per-request fees
- Cache reads at a discounted rate
Plans as listed on the official pricing page on . Prices can change and may vary by country or exclude tax. Check the official pricing page.
Perplexity Sonar API alternatives
The main Perplexity Sonar API alternatives listed on Debuly are Tavily, Exa, Brave Search API, You.com API, OpenRouter and Cohere. The table compares pricing and platforms; order is A–Z.
| Tool | Pricing | Starting price | Platforms |
|---|---|---|---|
| Perplexity Sonar API | Paid | — | API |
| Brave Search API | Paid | — | API, Web |
| Cohere | Freemium | — | API, Web, Private deployment |
| Exa | Freemium | — | API |
| OpenRouter | Freemium | — | Web, API |
| Tavily | Freemium | $30/mo | API, Web, Python, JavaScript |
| You.com API | Freemium | $5, usage-based | API, Web |

Brave Search API
Search API built on Brave's own web index, with LLM Context, web, news, image, video and place endpoints for apps and AI agents.

Cohere
Enterprise AI company offering Command, Embed and Rerank models by API, the North agent platform and private deployments.
FreemiumVisit Cohere (official site)
Exa
Web search API for AI agents and apps that returns relevant passages with citations, billed pay-as-you-go from a free monthly credit.
FreemiumVisit Exa (official site)
OpenRouter
Unified OpenAI-compatible API for 500+ AI models with provider fallbacks, pay-as-you-go credits and a free tier
FreemiumVisit OpenRouter (official site)
Tavily
Web search, extraction and crawling API built for AI agents and RAG, with a free tier of 1,000 credits a month.
FreemiumVisit Tavily (official site)
You.com API
You.com API provides real-time web search, page contents, cited answers and multi-step research endpoints for AI apps, billed per call.
Perplexity Sonar API FAQ
What is the Perplexity Sonar API?
It is Perplexity's API for web-grounded AI responses. Sonar provides chat completions through an OpenAI-compatible interface, and the wider platform adds the Agent, Search, Embeddings, Decisions and Orchestrator APIs.
How much does the Perplexity API cost?
It is pay as you go. The Search API is $5.00 per 1K requests ($1.00 with Fast Search), Agent API web search is $0.0025 per invocation, and embeddings start at $0.004 per 1M tokens. Model tokens are billed at each model's published rate.
Is there a free plan for the Perplexity API?
The pricing docs describe pay-as-you-go pricing with no subscription required, and they do not list a free tier; you pay for each request, tool call or million tokens used.
What plans does the Perplexity API have?
There are no subscription tiers. You pay per request, per tool invocation or per million tokens, depending on the API, using credits or auto top-off.
Is the Perplexity API compatible with OpenAI's SDK?
Yes. Sonar supports the OpenAI Chat Completions format, so you can use OpenAI client libraries with Perplexity's base URL and your API key.
Does the API provide citations?
Yes. The Agent API returns web-grounded answers with built-in citations in one call.
What are alternatives to the Perplexity Sonar API?
Alternatives include Tavily, Exa, Brave Search API, You.com API, OpenRouter and Cohere, depending on whether you need search results, grounded answers or model routing.
Sources & verification
Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.
Image credits
- Logo: from docs.perplexity.ai · fetched · Vendor trademark, used only to identify the product; removed on request
- Cover: none collected yet, so a generated tile is shown.
Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.



