← Back to docs

Batch API status roadmap

Current capabilities and upcoming batch processing endpoints.

Today: client concurrency

Meridian supports parallel requests from your client. Use Promise.all or your language's concurrency primitives to fire multiple /v1/analyze calls simultaneously. Each request is independently authenticated and metered.

const results = await Promise.all(
  ids.map(id => fetch(`/v1/analyze/${id}`))
);

Q3 2026: native /v1/batches

A single endpoint accepting an array of inputs, returning a batch ID for polling. Designed for high-throughput workflows with automatic retries, webhook callbacks, and consolidated billing.

  • Submit up to 1,000 items per batch
  • Poll GET /v1/batches/:id for completion status
  • Webhook POST on batch completion
  • Results available for 72 hours

Questions? Join our Discord for early access.