API · Limits

Rate limits & credit costs

Every key is rate-limited per minute. Discovery, strategy creation, trading, and signal calls are free; backtests and ML runs are metered.

Rate limits

Each key has a per-minute limit (default 60 requests/min), enforced as a sliding 60-second window. Exceeding it returns 429 RATE_LIMITED with Retry-After and RateLimit-* headers:

429 response headers
Retry-After: 60
RateLimit-Limit: 60
RateLimit-Remaining: 0

GET /api/v1/whoami reports your current window usage (used / remaining) so you can pace requests.

Credit cost per endpoint

Your own agent does not spend Trigr credits for reasoning, catalog search, or saving a draft. Trigr-hosted backtest and ML compute keeps the same credit policy as Studio. A pre-check returns 402 before an unaffordable run starts.

EndpointCredit cost
GET /api/v1/whoamiFree
GET /api/v1/catalogFree · public
POST /api/v1/graph-strategiesFree · strategy quota applies
POST /api/v1/ml/backtest50 credits on success + any signed provider usage
POST /api/v1/ml/optimizeML coverage + compute policy
POST /api/v1/strategiesComing soon
POST /api/v1/positionsComing soon
PATCH /api/v1/positions/{id}Coming soon
POST /api/v1/positions/{id}/closeComing soon
GET /api/v1/positionsComing soon
GET /api/v1/strategies/{id}/perfComing soon
GET /api/v1/signals/strategiesFree (Pro)
GET /api/v1/strategies/{id}/backtestFree (Pro)
GET /api/v1/strategies/{id}/liveFree (Pro)
GET /api/v1/strategies/{id}/signalFree (Pro)
GET /api/v1/signalsFree (Pro)

How credits are priced

  • 1,000 credits = $1. You're billed the real model cost, rounded up to at least 1 credit per call.
  • Cache hits cost nothing. Re-running an identical ML call that the service already computed doesn't re-charge.
  • 402 on empty balance. The pre-check blocks the run and returns the credits needed vs your balance.
402 INSUFFICIENT_CREDITS
{ "error": "INSUFFICIENT_CREDITS", "needed": 200, "balance": 40 }