Perplexity Sonar API
"Power your products with unparalleled real-time, web-wide research and Q&A capabilities." [1]
Perplexity Sonar API is a REST-based search and language model service that lets developers add real-time web retrieval, Q&A with inline citations, deep research, and semantic embeddings to their products. Pricing is usage-based starting at $1.00 per million tokens, with self-serve signup and no sales call required. The API ships SDKs for Python, TypeScript, and Node.js MCP, holds SOC 2 Type 2 certification, enforces a zero data-retention policy, and counts Samsung among its documented customers.
Best for / Avoid if
Best for: Regulated or enterprise workloads - compliance attestations and an enterprise plan; AI agents and automation - an agent-ready surface (MCP / llms.txt); Teams needing broad API coverage out of the box
Avoid if: You want to try it free before paying
Scores
- 90 / 100Agent friendliness
- 85 / 100Pricing transparency
- 60 / 100Setup speed
- 90 / 100Docs quality
- 85 / 100Procurement ease
- 25 / 100Trust readiness
Pricing & procurement
- Pricing model
- Usage-based
- Published pricing
- ✓ Yes
- Free tier
- ✗ No
- Self-serve signup
- ✓ Yes
- Requires sales call
- ✗ No
- Enterprise plan
- ✓ Yes [2]
| Plan | Item | Per | Amount | Source |
|---|---|---|---|---|
| Sonar | input tokens | 1M tokens | $1 | source |
| Sonar | output tokens | 1M tokens | $1 | source |
| Sonar | request fee (low search context) | 1,000 requests | $5 | source |
| Sonar | request fee (medium search context) | 1,000 requests | $8 | source |
| Sonar | request fee (high search context) | 1,000 requests | $12 | source |
| Sonar Pro | input tokens | 1M tokens | $3 | source |
| Sonar Pro | output tokens | 1M tokens | $15 | source |
| Sonar Pro | request fee (low search context) | 1,000 requests | $6 | source |
| Sonar Pro | request fee (medium search context) | 1,000 requests | $10 | source |
| Sonar Pro | request fee (high search context) | 1,000 requests | $14 | source |
| Sonar Pro | Pro Search request fee (fast/standard, low context) | 1,000 requests | $6 | source |
| Sonar Pro | Pro Search request fee (fast/standard, medium context) | 1,000 requests | $10 | source |
| Sonar Pro | Pro Search request fee (fast/standard, high context) | 1,000 requests | $14 | source |
| Sonar Pro | Pro Search request fee (pro/multi-step, low context) | 1,000 requests | $14 | source |
| Sonar Pro | Pro Search request fee (pro/multi-step, medium context) | 1,000 requests | $18 | source |
| Sonar Pro | Pro Search request fee (pro/multi-step, high context) | 1,000 requests | $22 | source |
| Sonar Reasoning Pro | input tokens | 1M tokens | $2 | source |
| Sonar Reasoning Pro | output tokens | 1M tokens | $8 | source |
| Sonar Reasoning Pro | request fee (low search context) | 1,000 requests | $6 | source |
| Sonar Reasoning Pro | request fee (medium search context) | 1,000 requests | $10 | source |
| Sonar Reasoning Pro | request fee (high search context) | 1,000 requests | $14 | source |
| Sonar Deep Research | input tokens | 1M tokens | $2 | source |
| Sonar Deep Research | output tokens | 1M tokens | $8 | source |
| Sonar Deep Research | citation tokens | 1M tokens | $2 | source |
| Sonar Deep Research | reasoning tokens | 1M tokens | $3 | source |
| Sonar Deep Research | search queries | 1,000 queries | $5 | source |
| Search API | web search requests | 1,000 requests | $5 | source |
| Embeddings API | pplx-embed-v1-0.6b tokens | 1M tokens | $0.004 | source |
| Embeddings API | pplx-embed-v1-4b tokens | 1M tokens | $0.03 | source |
| Embeddings API | pplx-embed-context-v1-0.6b (contextualized) tokens | 1M tokens | $0.008 | source |
| Embeddings API | pplx-embed-context-v1-4b (contextualized) tokens | 1M tokens | $0.05 | source |
| Agent API | web_search tool call | call | $0.005 | source |
| Agent API | fetch_url tool call | call | $0.0005 | source |
| Agent API | people_search tool call | call | $0.005 | source |
| Agent API | finance_search tool call | call | $0.005 | source |
| Agent API | sandbox session (20-minute window) | session | $0.03 | source |
Capabilities
- Supported actions
- chat_completions_with_web_search (POST /v1/chat/completions), async_chat_completions (POST /v1/async/sonar), get_async_response (GET /v1/async/sonar/{api_request}), list_async_requests (GET /v1/async/sonar), web_search (POST /v1/search), embeddings (POST /v1/embeddings), contextualized_embeddings (POST /v1/embeddings/contextualized), agent_completions (POST /v1/agent), list_models (GET /v1/models)
- Regions
- Global web index; compute hosted on AWS in North America [3]
- Languages
- Queries accepted in multiple languages via ISO 639-1 search_language_filter parameter (up to 10 languages per request); English is primary
- Input types
- natural language query, structured chat messages (OpenAI-compatible), domain filter arrays, date/recency filters, ISO 639-1 language filter, ISO 3166 country/location filter, base64-encoded images, HTTPS image URLs
- Output types
- JSON (OpenAI-compatible chat completion), streaming server-sent events, citations array with source URLs, search_results metadata, token usage metrics, structured JSON schema output, vector embeddings
- Webhooks
- ✗ No
- Sandbox / test mode
- ✗ No
- SDK languages
- Python, TypeScript, MCP (Node.js) [4]
- MCP server
- ✓ Yes [5]
Trust & compliance
- SOC 2
- SOC 2 Type II [6]
- HIPAA
- ✗ No [7]
- GDPR
- – Unknown
- ISO 27001
- – Unknown
- PCI DSS
- – Unknown
- Published SLA
- ✗ No [8]
- Rate limits
- Tier-based RPM by model. Sonar / Sonar Pro / Sonar Reasoning Pro: Tier 0=50 RPM, Tier 1=150, Tier 2=500, Tier 3=1,000, Tier 4=4,000, Tier 5=4,000. Sonar Deep Research: Tier 0=5, Tier 1=10, Tier 2=20, Tier 3=40, Tier 4=60, Tier 5=100 RPM. Search API: 50 req/s fixed across all tiers. Tiers advance by cumulative lifetime spend: Tier 1=$50+, Tier 2=$250+, Tier 3=$500+, Tier 4=$1,000+, Tier 5=$5,000+. [9]
- Known restrictions
- Zero data retention policy: no customer prompt/response data is retained after request processing, Customer data is never used for model training, Protected Health Information cannot be processed without a signed Business Associate Agreement, Fine-tuning of models is not supported, No uptime SLA or service quality guarantees published ('We do not guarantee this at the moment'), SafeSearch content filtering is on by default, Domain and date filters cannot be combined in certain ways; max 20 domains per filter, Perplexity may suspend access for applications deemed competitive with Perplexity, Search API Terms of Service explicitly states no enhanced security obligations or uptime guarantees apply to Search Services
Developer surface
Integration
Adoption & maturity
- Launched
- 2023-10-04
- GA
- 2025-01-01
- Notable customers
- Samsung
Other Web Search APIs
Firecrawl
The API to search, scrape, and interact with the web at scale.
Exa
"Web search, built for AI agents. One API for search, crawling, and research agents."
Brave Search API
"The Brave Search API provides developers with access to Brave Search results and enables building AI products, chatbots, coding assistants, and AI-search engines with real-time web data."
Tavily Search API
"Real-time search, extraction, research, and web crawling through a single, secure API."
Linkup
"Production-grade Web Search API for AI"
Jina AI Search Foundation (DeepSearch / Reader)
"Your Search Foundation, Supercharged."
References
- ↑Description: docs.perplexity.ai
- ↑Enterprise plan: perplexity.ai
- ↑Regions: docs.perplexity.ai
- ↑SDK languages: docs.perplexity.ai
- ↑MCP server: github.com · docs.perplexity.ai
- ↑SOC 2: docs.perplexity.ai
- ↑HIPAA: docs.perplexity.ai · hub-prod.perplexity.ai
- ↑Published SLA: docs.perplexity.ai · perplexity.ai
- ↑Rate limits: docs.perplexity.ai · docs.perplexity.ai
Change history
- 2026-06-15 Score Agent Friendliness: 60 → 90
- 2026-06-15 Score Docs Quality: 55 → 90
- 2026-06-14 Robots Allows Agents: (none) → Yes
- 2026-06-14 Markdown Docs URL: (none) → https://docs.perplexity.ai/docs.md
- 2026-06-14 Has Structured Data: (none) → Yes
- 2026-06-14 API Reference URL: (none) → https://docs.perplexity.ai/api-reference/agent-post
- 2026-06-14 Markdown Docs Served: (none) → Yes
- 2026-06-14 Capabilities: {} → {"semantic":true,"ai_answer":true,"agent_ready":true,"zero_retention":true}
- 2026-06-14 Summary Md: (none) → Perplexity Sonar API is a REST-based search and language model service that let…
- 2026-06-14 Best For: (none) → Regulated or enterprise workloads - compliance attestations and an enterprise p…
- 2026-06-14 Score Agent Friendliness: (none) → 60
- 2026-06-14 Score Pricing Transparency: (none) → 85
- 2026-06-14 Score Setup Speed: (none) → 60
- 2026-06-14 Score Docs Quality: (none) → 55
- 2026-06-14 Score Procurement Friction: (none) → 85
- 2026-06-14 Score Trust Readiness: (none) → 25
- 2026-06-14 Avoid If: (none) → You want to try it free before paying
- 2026-06-14 Scoring Methodology: (none) → Scores are computed deterministically from this profile's published, sourced fi…
- 2026-06-14 Llms Txt Present: (none) → Yes
- 2026-06-14 Docs URL: (none) → https://docs.perplexity.ai/docs
- 2026-06-14 Changelog URL: (none) → https://docs.perplexity.ai/changelog
- 2026-06-14 Openapi Spec URL: (none) → https://docs.perplexity.ai/openapi.json
- 2026-06-14 Rendering: (none) → static
- 2026-06-14 Llms Txt URL: (none) → https://docs.perplexity.ai/llms.txt
- 2026-06-14 SLA Published: set to No
- 2026-06-14 Data Retention Policy URL: set to https://docs.perplexity.ai/docs/resources/privacy-security
- 2026-06-14 Documented Rate Limits: set to Tier-based RPM by model. Sonar / Sonar Pro / Sonar Reasoning Pro: Tier 0=50 RPM…
- 2026-06-14 Rate Limit Requests: set to 50
- 2026-06-14 Rate Limit Window: set to minute
- 2026-06-14 Known Restrictions: set to Zero data retention policy: no customer prompt/response data is retained after …
- 2026-06-14 Auth Methods: set to api_key
- 2026-06-14 Auth Docs URL: set to https://docs.perplexity.ai/docs/getting-started/quickstart
- 2026-06-14 API Style: set to rest
- 2026-06-14 Base URL: set to https://api.perplexity.ai
- 2026-06-14 API Version: set to v1
- 2026-06-14 Versioning Scheme: set to url
- 2026-06-14 Stability: set to ga
- 2026-06-14 Deprecation Policy URL: set to https://docs.perplexity.ai/changelog/changelog
- 2026-06-14 MCP URL: set to https://docs.perplexity.ai/guides/mcp-server
- 2026-06-14 Quickstart URL: set to https://docs.perplexity.ai/docs/getting-started/quickstart
- 2026-06-14 Idempotency Supported: set to No
- 2026-06-14 Error Format: set to vendor-specific
- 2026-06-14 Requires Verification: set to No
- 2026-06-14 Starting Price Usd: set to 1
- 2026-06-14 Price Basis: set to 1M tokens
- 2026-06-14 Launched At: set to 2023-10-04
- 2026-06-14 GA Date: set to 2025-01-01
- 2026-06-14 Slug: set to perplexity-sonar
- 2026-06-14 Fields Not Found: set to iso_27001, gdpr, pci_dss, free_tier_details, minimum_commitment, context_window…
- 2026-06-14 Source Confidence: set to high
Suggest an edit / leave a review
Leave a review or comment
curl -X POST https://apio.sh/api/feedback/perplexity-sonar \
-H 'Content-Type: application/json' \
-d '{"kind":"review","rating":5,"body":"Your experience with this API…"}'Suggest a correction to a field (cite a source)
curl -X POST https://apio.sh/api/suggest/perplexity-sonar/FIELD \
-H 'Content-Type: application/json' \
-d '{"value":"corrected value","citations":[{"url":"https://source.example/page","excerpt":"supporting quote"}],"note":"what changed and why"}'