Perplexity APIAsk a natural-language research question with one SocialCrawl API key. Get a written answer plus source URLs as structured JSON. Powered by Perplexity Sonar through the AI gateway at 1 credit per live call. No Perplexity developer account.
1 active endpoint. Cited answer plus sources.
One live research endpoint. Natural-language query in, synthesised answer plus citation URLs out. Best for open questions that need current web grounding, not a fixed social-platform lookup. Pair with Tavily and search/everywhere when you also need ranked hits or social fan-out.
/v1/perplexity/researchNatural-language research over the live web via Perplexity Sonar. Returns data.answer plus data.sources as [{ url, title? }]. Sources may be empty on short factual replies. Auto-refund on upstream failure.
query
Returns a written answer to a natural-language question researched over the live web, along with the source URLs the answer was built from.
Use it for open questions that need current information, such as recent news or funding, rather than a fixed lookup on one platform.
query · Natural-language research prompt. Sonar autonomously searches the live web and grounds the response in real sources. No prompt-engineering required — phrase it as you would to a search engine or research assistant.
$ curl https://www.socialcrawl.dev/v1/perplexity/research?query=What+is+the+capital+of+France%3F \
-H "x-api-key: sc_YOUR_API_KEY"// Running this live needs your own API key — hit "Try it" to see an example responsePerplexity is a normal SocialCrawl surface under /v1/perplexity. You call GET with an API key, spend 1 credit on a live miss, and get the shared JSON envelope. Sonar runs through the AI gateway. No Perplexity OAuth and no second SDK.
Send your key in the x-api-key header. No Perplexity developer account, no separate Sonar key. The same SocialCrawl key works for Tavily, search/everywhere, and the social catalog.
The research route is GET. Pass query as a natural-language prompt. We validate required params before charging.
Live research costs 1 credit. Cache hits cost 0. Empty or hard failures refund. Bad params return 400 with no charge. Insufficient balance returns 402 with no charge.
Every response is the same shape: success, data, credits_used, credits_remaining, request_id, cached. data holds answer and sources[].
Most products start with a grounded answer, then deepen with web search or social fan-out when they need ranked URLs or platform-native hits.
GET /v1/perplexity/research?query=…Analytics: answer plus sources[{ url, title? }]
Freeform questions that need live web grounding land here first.
GET /v1/tavily/search?query=…Ranked web hits and optional extract bodies
When you need a result list or full page text, not only a synthesised reply.
GET /v1/search/everywhere?query=…search/everywhere: multi-platform social hits
When the question is what people are saying on social, not only the open web.
GET /v1/perplexity/research
?query=what+is+rrf+fusion+in+search
Host: www.socialcrawl.dev
x-api-key: sc_your_api_key_here{
"success": true,
"data": {
"answer": "…",
"sources": [
{ "url": "https://example.com", "title": "…" }
]
},
"credits_used": 1,
"credits_remaining": 9999,
"request_id": "req_…",
"cached": false
}Archetype is Analytics so the envelope keeps answer and sources instead of forcing a social Post shape.
answer: synthesised natural-language reply grounded in live web sources
sources[]: { url, title? }; may be empty for short factual answers
Same request lifecycle as every other /v1 platform endpoint. Perplexity is not a sidecar product.
Next.js catch-all routes into the Hono social API. We mint a request_id, authenticate the key, then enforce rate limit and concurrency for the key.
Registry lookup finds perplexity/research. Required query validation runs first. Invalid input returns 400 with no charge. Valid calls debit 1 credit atomically before upstream work.
A deterministic cache key is built from platform, resource, and params. Hit: return immediately with credits_used = 0. Miss: invoke Perplexity Sonar through the Vercel AI gateway. Retries on 5xx and network errors where the policy allows.
Upstream text and citations map to Analytics { answer, sources[] }, validate against the envelope schema, wrap success metadata, and log for billing audit.
Billing rules that matter in production
Perplexity research is live-web grounding through Sonar. We expose the answer and citations under the shared SocialCrawl envelope so your code does not open a second vendor account.
Natural-language research over the open web. Returns a written answer and the URLs used as evidence. Not a social feed, profile, or comment tree endpoint.
Upstream is Perplexity Sonar via the Vercel AI gateway (kind ai-perplexity). Not ScrapeCreators. No customer-side Perplexity key or developer signup.
A unified JSON envelope: success, data, credits_used, request_id, cached. data is Analytics-shaped with answer and sources[{ url, title? }].
No chat history, no multi-turn sessions, no write tools. No separate Perplexity billing surface. For ranked SERP rows or page extract use Tavily or Web; for social fan-out use search/everywhere.
The jobs this API is most often used for.
Grounded research answers with citations
Callers use research for open questions that need current web evidence: news, funding, product facts, and short research memos. Many pipelines then call Tavily or search/everywhere for ranked lists or social context. Cited Sonar answers at 1 credit with no Perplexity developer account. Same SocialCrawl key as Tavily and search/everywhere.
Live research depends on Sonar and the AI gateway. Expect longer than a simple social profile read; treat it as a research call, not a sub-second cache fill.
Common ways teams put this data to work, and the stack each one tends to run.
Turn freeform product or market questions into answer plus sources for memos, RAG stubs, and internal bots.
Pull a grounded draft answer and citation URLs before deeper SERP or social sweeps with Tavily and search/everywhere.
Wire research into pipelines that already use the SocialCrawl key for social and web endpoints. Cache hits keep exact repeats cheap.
Research is 1 credit per live call. Cache hits are free. Pair with Tavily or search/everywhere when you need more than a cited answer.
curl "https://www.socialcrawl.dev/v1/perplexity/research?query=what+is+rrf+fusion+in+search" \
-H "x-api-key: sc_your_api_key_here"Same key as the rest of the catalog
Every endpoint returns structured JSON in a unified envelope. Computed fields such as engagement rate and content category are included only where the endpoint supports them and the required source inputs are present.
Same Perplexity data, two very different paths. Auth, rate limits, schema, and cost — side by side.
| Feature | SocialCrawl | Perplexity API (direct) |
|---|---|---|
| Authentication | One x-api-key, the same key as your social and dev data | Separate Perplexity developer account and API key |
| Setup | Single GET request with a query parameter | Chat-completions integration with model and parameter management |
| Pricing | 1 credit per research call, one balance across 48 platforms | Separate token-based billing you fund and track independently |
| Response schema | Cited answer in the unified SocialCrawl JSON envelope | OpenAI-style chat completion payload |
| Data coverage | Sonar research plus Tavily, social search, and dev data on one key | Full Sonar model family with fine-grained model control |
| Free tier | 100 free credits on signup, no credit card | Prepaid credits; cost varies by model and tokens |
Authentication
Setup
Pricing
Response schema
Data coverage
Free tier
Find answers to frequently asked questions about SocialCrawl's API, pricing, and capabilities.
Contact usAsk AI about SocialCrawl