Developers
Aetherya is an HTTP API with a companion MCP connector. Everything the dashboard does is available over the same surface, scoped to a workspace and billed against the same plan.
Authentication
Send an organization API key as a bearer token:
curl https://api.aetherya.ai/api/usage/status \ -H "Authorization: Bearer aeth_live_..." \ -H "X-Aetherya-Workspace-Id: <workspace-uuid>"
Organization keys are minted by an owner or admin under Settings → Security. They are owned by the organization rather than by a person, so they survive that person leaving, and they can carry an expiry. A key may also carry a default workspace, in which case the X-Aetherya-Workspace-Id header is optional — a header you do send always wins.
Rate limits
Limits are enforced per API server replica, so the effective ceiling is the figure above multiplied by the number of replicas serving traffic (currently up to two). Over the limit, the API answers 429 with a retryAfterSeconds field.
Endpoints
Hand-written specification covering Aetherya's public HTTP surface: pricing, usage and estimation, cognitive audits, personas, Ad Lab runs and Audience Chat. It is maintained alongside the API rather than generated from it, so treat an endpoint that behaves differently from this document as a bug in the document and tell us.
General
Get canonical pricing tiers
Get approved API-equivalent usage values
Get percentage-based usage status
Estimate a workload as a percentage range
Get Business auto-reload settings
Update Business auto-reload settings
Start a behavioral-simulation audit of a URL
Poll the status of an audit job
Fetch a completed audit report
List synthetic audiences (personas)
Generate synthetic audience personas
Test copy or messaging variants with a synthetic audience
Create an audience chat session
Send a message to a synthetic audience chat session
Full specification: openapi.json.
MCP connector
For agents, @aetherya/mcp exposes the same capabilities as tools — planning a study, running it, collecting the result and turning it into a decision brief — without writing HTTP calls. Setup and tool reference.
Availability and support commitments are in the SLA; current platform health is on the status page.