API rate limit calculator
Describe the job — how many items, how many calls each, how often it runs — and find out the pace it needs and whether the published limits allow it.
GET /v1/satellite/creator/:platform/:username
- Calls per pass
- 500
- Calls per day
- 500
- Published limits
- 5 / min · 100 / hour · 1,000 / day
- Pace to stay under
- 1 call every 12.0s
- Shortest pass
- 1h 40m
- Daily headroom
- 500 calls left
On a 429, wait the number of seconds in Retry-After rather than guessing; if it is absent, back off 1s, 2s, 4s, 8s, 16s, 32s with jitter. Every response carries your remaining budget in the rate-limit headers. Full table: /docs/rate-limits.
Three numbers decide everything
Items, calls per item, and passes per day. Multiply the first two and you have the size of one pass; multiply by the third and you have the daily total. Compare the pass against the per-minute cap and the total against the daily cap, and you know whether the design works before you write it.
The constraint that catches people is not the daily ceiling. It is wall-clock time: at a 5-per-minute cap, a thousand-item pass needs more than three hours no matter how many workers you run, because the cap is on the key and not on the process.
What to do when it does not fit
- Batch. A batch creator lookup takes up to 25 creators in one call and counts against the 10,000-per-day limit instead of the 5-per-minute one.
- Stop polling. Replace a status loop with a webhook. Most jobs that do not fit are mostly polling.
- Cache. A creator answered from the six-hour cache costs nothing and takes no quota.
- Spread it. A tracking sweep rarely needs every item at the same hour.
- Ask. Custom limits exist, and the contact form on the pricing page reaches the team.
Read the headers, not the clock
Every response carries the budget remaining for whichever window is closest to running out, so the same endpoint can report a per-minute figure on one call and a per-hour figure on the next. Log those headers from day one: they turn a future 429 from a mystery into a number you already had.
FAQ
- How do I convert a per-minute limit into a delay between calls?
- Divide 60 by the limit. A 5-per-minute cap is one call every 12 seconds; a 50-per-minute cap is one call every 1.2 seconds. Pacing to that number is steadier than bursting and then waiting, and it keeps your remaining budget predictable.
- What are the Virlo rate limits?
- Creator lookups and video outlier checks are 5 per minute, 100 per hour and 1,000 per day. Hashtag stats, video digests and trends are 50 per minute, 500 per hour and 5,000 per day. Emerging trends is 60 per minute. Agents: reading an agent and its results is 500 per minute, other agent reads 120, and creating an agent or asking for keyword suggestions 30. Everything else has no per-minute cap and a 10,000-per-day ceiling. Each endpoint keeps its own counter.
- What should I do when I get a 429?
- Wait the number of seconds in the Retry-After header. It is the server telling you exactly when the window resets, so it beats any backoff you would guess. Only when it is absent should you fall back on exponential backoff with jitter, which stops a fleet of workers retrying in lockstep.
- Why does my job fail the check even though it is under the daily cap?
- Because a pass has to finish before the next one starts. At 5 calls per minute, 1,000 items take over three hours, so an hourly job overlaps itself while still being well inside the daily allowance. Spread the work or run fewer passes.
- Can I raise the limits?
- Yes. Custom limits are available on request, and the headers show yours rather than the defaults once they are applied. Batch endpoints also help without any change: a batch creator lookup takes up to 25 creators per call and bills per creator rather than per call.
- How many requests should I run in parallel?
- Enough to keep the pace, not more. With a per-minute cap, concurrency buys nothing beyond what the cap allows and just converts into 429s. On the uncapped endpoints the useful limit is your own backend and the latency you are willing to absorb.
Get the data behind the tool
The Virlo API returns TikTok, Instagram Reels and YouTube Shorts data by the call — hashtags, videos, sounds, creators and trends. Pay per call, no scrapers to maintain.
Create your API key