What Causes Perplexity Rate Limits
Perplexity applies rate limits in two distinct ways:
1. Per-Minute Throttle (All Users)
If you send too many searches in a short window, Perplexity temporarily blocks further requests. This is a traffic-shaping mechanism to prevent server overload and applies to both free and Pro users.
2. Daily Pro Search Cap (300/day for Pro Users)
Pro users get 300 Pro searches per day. These are searches that use premium models (GPT-4o, Claude, Gemini, Sonar Large). Once you hit 300, Pro model access is blocked until the daily reset.
| Limit Type | Who It Affects | Reset Time | |------------|---------------|------------| | Per-minute throttle | All users | 1–5 minutes | | Daily Pro search cap | Pro users | Midnight UTC | | Free tier daily limit | Free users | Midnight UTC |
How to Identify Which Limit You Hit
Per-minute throttle:
- Error appears after rapid consecutive searches
- Clears in a few minutes
- Message: "Too many requests" or "Rate limit exceeded"
Daily Pro search cap:
- Error appears after heavy use that day
- Shows "Pro search limit reached" or "Upgrade for more searches"
- Won't clear until midnight UTC
Quick search vs Pro search:
- Quick search (default/Sonar model) is unlimited for Pro users — if you can still use that, you've likely hit only the Pro model cap
Step-by-Step Fix
1. Switch to Quick Search (Immediate Fix)
If you've hit the Pro search cap or a per-minute throttle, Quick search continues to work. For most research tasks, Quick search delivers good results.
How to switch:
- Click the model selector at the top of the search box
- Select Default or Sonar instead of a premium model
Quick searches (unlimited on Pro) handle factual lookups, summaries, comparisons, and most everyday research.
2. Wait Out the Per-Minute Throttle
If you've been sending rapid-fire searches:
- Stop searching entirely for 60–90 seconds
- Do not refresh or retry during this window
- Resume with a slower pace (one search every 5–10 seconds at minimum)
3. Check Your Daily Reset Time
The daily cap resets at midnight UTC. To find your local reset time:
| Timezone | Daily Reset (Local Time) | |----------|--------------------------| | UTC-8 (US Pacific) | 4:00 PM | | UTC-5 (US Eastern) | 7:00 PM | | UTC+0 (UK) | Midnight | | UTC+1 (Central Europe) | 1:00 AM next day | | UTC+8 (China / Singapore) | 8:00 AM next morning |
4. Reduce Search Frequency
Avoid these behaviors that trigger rate limits quickly:
- Sending follow-up searches less than 2–3 seconds apart
- Opening multiple browser tabs and searching simultaneously
- Using browser automation or scraping tools against the web UI
5. Upgrade to Pro (If on Free Tier)
Free users have tighter daily limits. Perplexity Pro gives you:
- 300 Pro searches per day
- Unlimited Quick searches
- Access to GPT-4o, Claude, Gemini, Sonar Large
Upgrade at perplexity.ai/pro.
6. Use the Perplexity API for High-Volume Use
If you're building a product or doing automated research, the web app is the wrong tool. The Perplexity API provides:
- Separate quota from your app account
- Programmatic rate limit handling
- Higher throughput at paid API tiers
API access and pricing are available at docs.perplexity.ai.
API Rate Limits (For Developers)
If you're using the Perplexity API (HTTP) and seeing 429 errors:
- Check your API tier limits in your developer dashboard
- Implement exponential backoff — wait 1s, then 2s, then 4s on each retry
- Reduce parallelism — avoid running more than 2–3 concurrent API requests unless your tier explicitly supports it
- Monitor token usage — large prompts consume more tokens per minute (TPM)
App limits and API limits are fully separate — using one does not affect the other.
Common Mistakes to Avoid
- Retrying immediately after a 429 error — this makes the throttle worse and delays your recovery
- Running multiple browser tabs — each tab counts as separate requests against the same rate limit
- Confusing per-minute throttle with daily cap — a short wait fixes the throttle; only the UTC reset fixes the daily cap
- Using the web app for automation — the web UI is not designed for programmatic use; use the API instead
- Expecting unlimited Pro searches — Quick search is unlimited, but Pro searches (premium models) are capped at 300/day
May 2026 Context: What Has Changed With Perplexity Limits
As of May 2026, Perplexity has made several adjustments to how rate limits are enforced that are worth understanding:
Pro search cap remains 300/day. The core daily cap for Pro subscribers has not changed since early 2025. The 300 Pro searches per day limit applies regardless of which premium model you select — GPT-4o, Claude, Gemini, and Sonar Large all count equally against this cap.
Deep Research counts more heavily. Perplexity's Deep Research feature, which runs multi-step research agents, deducts more than one Pro search credit per run. Deep Research reports can consume 3–5 Pro search credits per session depending on the complexity and number of sources retrieved. If you use Deep Research regularly, your 300/day cap can deplete much faster than simple Pro searches.
Per-minute throttle has tightened. User reports from early 2026 indicate that the per-minute throttle is more strictly enforced than it was in 2024. Sending more than approximately 8–10 searches within a 60-second window now reliably triggers a temporary block. The recommended pace for sustained research sessions is one search every 5–8 seconds.
Spaces and Collections features. Perplexity added Spaces (persistent research collections) in late 2025. Searches run inside a Space still count against the same rate limits as regular searches — there is no separate quota for Space-based research.
API vs. Web App Rate Limits: Key Differences
If you use both the Perplexity web app and the API, understanding how limits are separated is important:
| Dimension | Web App (Pro) | API | |-----------|---------------|-----| | Daily Pro search cap | 300/day | Not applicable (billed per token) | | Per-minute throttle | ~8–10 searches/min | Tier-dependent (RPM in dashboard) | | Quota tracking | Resets at midnight UTC | Rolling window (per API tier) | | Usage counting | Pro model usage only | All requests count | | How to check | perplexity.ai/settings | docs.perplexity.ai dashboard |
Critical distinction: Using the API does not draw from your 300 daily Pro searches. Conversely, using the web app does not affect your API rate limits. They are completely independent systems on separate billing and quota infrastructure.
When to use the API instead of the web app:
- You need more than 300 research queries per day
- You want to automate or batch process queries
- You need programmatic access to responses (JSON format)
- You are building a product or workflow on top of Perplexity's capabilities
API 429 errors (Too Many Requests): If you hit a 429 from the API, implement exponential backoff — wait 1 second, retry; if it fails again, wait 2 seconds, then 4 seconds. Never retry immediately after a 429, as this extends the throttle period.