Token-based limits
Every authenticated request costs tokens. Your tier defines how many tokens you can spend per second. Most requests cost the default of 10 tokens; each operation’s API reference page lists its cost where it differs (order cancellations, single-order reads, quote create/cancel, and multivariate-collection lookup are currently cheaper than the default).
Your effective rate for any given endpoint is budget ÷ cost.
Reads and writes are billed separately
You have two independent token budgets:
| Bucket | What it covers |
|---|
| Read | GET endpoints and anything not explicitly routed elsewhere. |
| Write | Order placement, amends, cancels, order groups, and the RFQ quote flow. |
Batch endpoints don’t save tokens
A batch request costs the same as making each call individually — every item in the batch is billed separately:
Tiers and budgets
Per-second token budgets in each bucket:
| Tier | Read budget | Write budget |
|---|
| Basic | 200 | 100 |
| Advanced | 300 | 300 |
| Premier | 1,000 | 1,000 |
| Paragon | 2,000 | 2,000 |
| Prime | 4,000 | 4,000 |
Tier qualification
- Basic: complete account signup.
- Advanced: complete the Advanced API application.
- Premier, Paragon, Prime: qualification criteria will be published shortly.
Kalshi may, at its discretion, adjust your tier at any time — including downgrading you from higher tiers following prolonged inactivity. Members may request an upgrade by contacting support with a description of their use case.
When you hit the limit
A rate-limited request returns 429 Too Many Requests with the body:
{"error": "too many requests"}
Retry-After or X-RateLimit-* headers. Apply exponential backoff on 429 until your bucket refills.
Bursting above your per-second budget
Your Write bucket holds two seconds of your per-second budget — so idle or below-steady traffic builds up headroom you can spend in a single burst. Useful for event-driven clients reacting quickly to market moves or price prints. Each request drains the bucket by its token cost; the bucket refills continuously at your per-second budget up to its capacity.
Over-drawing returns 429 Too Many Requests. There’s no enforced cooldown — your next request is allowed as soon as the bucket has enough tokens to cover it.
Basic tier is the exception — its Write bucket holds one second of budget, with no accumulated headroom.