Documentation Index
Fetch the complete documentation index at: https://mintlify.com/sagar-grv/ayush-synapse/llms.txt
Use this file to discover all available pages before exploring further.
Rate limiting is handled by Flask-Limiter, keyed by the client’s remote IP address. Limits are applied globally to all endpoints, with stricter per-route overrides on computationally intensive operations.
Default Limits
| Scope | Limit |
|---|
| All endpoints (global default) | 200 requests / day |
| All endpoints (global default) | 50 requests / hour |
GET /icd11/search/{term} | 30 requests / minute |
POST /translate | 30 requests / minute |
POST /bundle | 20 requests / minute |
POST /translate counts against the 30/minute route limit and the 200/day / 50/hour global limits simultaneously.
Flask-Limiter attaches the following headers to every response so clients can track their current standing:
| Header | Description |
|---|
X-RateLimit-Limit | The maximum number of requests permitted in the current window |
X-RateLimit-Remaining | The number of requests remaining before the limit resets |
X-RateLimit-Reset | Unix timestamp (UTC) at which the current window resets |
When a Limit Is Exceeded
When a client surpasses any limit, the API responds with:
HTTP/1.1 429 Too Many Requests
{
"error": "Too Many Requests"
}
X-RateLimit-* headers so the client knows when it can retry.
Configuring Limits
Rate limit storage is controlled by the RATELIMIT_STORAGE_URL environment variable.
| Value | Behaviour |
|---|
memory:// (default) | Counters are stored in process memory. Limits reset on every server restart. Each worker process maintains independent counters — not suitable for multi-worker deployments. |
redis://host:port/db | Counters are stored in Redis and shared across all processes and replicas. Recommended for production. |
.env file or deployment environment:
# Development (default)
RATELIMIT_STORAGE_URL=memory://
# Production — shared Redis instance
RATELIMIT_STORAGE_URL=redis://redis-host:6379/0
During local development and unit testing, the in-memory backend is sufficient and requires no additional infrastructure. Rate limits are rarely hit in a single-developer workflow.For load testing or high-volume integration scenarios — for example, bulk code translation or repeated FHIR Bundle ingestion — configure a Redis backend and coordinate request pacing with your deployment team to avoid unintended 429 responses.