It’s that time of the month when Perplexity AI rolls out some pretty slick updates that make your AI experience smoother and smarter. This September, things have gotten even better with fresh additions like a brand-new Search API built for real-time, sharp-as-a-tack web data, and more accessible structured outputs for developers. Whether you’re coding a chatbot, crafting marketing campaigns, or simply exploring AI’s potential, these updates make it easier and faster to get reliable, relevant results.
—
✅ Perplexity Launches Search API for Real-Time Web Data
Perplexity has introduced a new Search API that taps into its massive infrastructure behind the scenes, allowing developers to get fresh, real-time search results from hundreds of billions of webpages. Unlike older APIs that return summarized, synthesized answers, this one delivers fine-grained, ranked snippets straight from the web.
For developers, it means they can build AI applications or agents that fetch raw, accurate data without extra preprocessing. Marketers can use this API to sync inventory or generate campaign briefs using the freshest info. Analysts get faster access to up-to-date research materials, and content creators can pull citations with confidence knowing the data is fresh and accurate.
—
✅ Structured Outputs Now Open to All API Users
Perplexity took on board developer feedback and removed the previous premium tier restriction on structured outputs. Now all API users, no matter their subscription, can get JSON structured outputs across all models, with both JSON and Regex outputs supported for the advanced Sonar models.
This means that if you’re auto-summarising call transcripts or integrating responses into apps, your workflow just got a whole lot easier without hitting any paywalls. Regex support coming soon for all models will make pattern matching and data extraction even more seamless.
—
✅ Enhanced Sonar Models: New Search Modes & Simplified Pricing
The Sonar models powering Perplexity have seen upgrades to perform better and cost less. They now sport three new search modes:
- High: For deep, complex queries that need exhaustive answers.
- Medium: Balances detail and cost; great for everyday tasks.
- Low: Budget-friendly for straightforward, simple questions.
Developers and analysts can now pick the mode that suits their use case, whether it’s intensive research or quick lookups. Billing is simpler too, focusing on input/output tokens without charging for citation tokens (except deep research mode), helping build with clearer costs.
—
✅ Free Tier Updates & Quarterly Model Refreshes
For those just dipping their toes in Perplexity with the free plan, there’s good news and a wee bit of patience needed. The free tier offers solid daily AI access, providing up to 48,000 tokens per query and 50 queries per day with reasonable speed.
However, free users will continue to receive model upgrades on a quarterly schedule, with the next refresh due in October 2025, including updates to a large reasoning model. So while occasional advanced features and updates hit Pro and Enterprise subscribers first, casual users still get a reliable, supported AI at no cost.
—
If you’ve ever found yourself wishing you could just grab pinpoint, current info straight from the web for a client pitch, or wished your AI assistant could spit out data that slots right into your CRM without a fuss, these updates are a big step forward. The balance Perplexity finds between powerful tech and open, practical access means you don’t need to be a full-time coder or data scientist to tap into next-gen AI tools.
So why not take these new features for a spin? Head on over to Perplexity AI, explore the latest enhancements, and if you fancy, drop your thoughts to help shape future updates. Staying on top of these developments can give your workflows that wee bit of edge on typical days when everything feels a bit too slow or outdated.
