API Reference
The Nexatron REST API is organized around resource groups. All endpoints are prefixed with /api/v1 and require authentication.
Base URL: https://api.nexatron.io/api/v1
Authentication: pass your API key in the Authorization header:
Authorization: Bearer nxa_your_api_key_here
API keys are issued from the in-product Settings → API tokens page. Keys are scoped to a single tenant and can be revoked at any time.
This page is a conceptual overview. The endpoint reference (request and response schemas, parameters, and status codes) is generated from the reviewed, version-pinned API specification. Browse it under API Reference. It describes that published contract, not an inspection of your running deployment.
Resource groups
For a first integration, start with the Python quickstart or JavaScript SDK quickstart, then use the schemas below when you need exact fields. Do not begin by implementing every endpoint.
- Authenticate with an identity scoped to the intended tenant.
- Send a known-answer question against an authorized connection.
- Inspect the returned result and handle authentication, validation, and rate-limit errors.
The API surface is grouped by the core concepts of the platform:
| Group | What it covers |
|---|---|
| Query | Submit a natural-language question and receive an answer, with the generated SQL, the sources used, and a trust label. Streaming is available via server-sent events. |
| Connections | Create, list, inspect, test, and remove data-source connections, and retrieve a connected source's schema. |
| Semantic layer | Define and manage approved metrics, dimensions, and relationships so every query resolves to approved business definitions. |
| Conversations | List, retrieve, and manage conversations and their messages for multi-turn querying. |
| Feedback | Record thumbs-up/down and chart feedback that feeds accuracy improvements. |
Refer to the generated reference for the exact paths, methods, and payloads in each group.
Data sources
Nexatron connects to warehouses and databases (PostgreSQL, MySQL, SQL Server, Snowflake, BigQuery, Databricks, Redshift, Amazon S3 and DuckDB), to Google Sheets and Microsoft Excel files, and to business applications such as Salesforce and QuickBooks through a separate application connector layer.
The set of connector types available to your tenant is returned by the connector-types endpoint rather than hard-coded here, so it always reflects what the platform currently supports. See GET /connections/types in the generated reference.
Common Response Codes
| Code | Meaning |
|---|---|
200 | Success |
201 | Created |
400 | Bad request (check the detail field) |
401 | Invalid or missing API key |
403 | Insufficient permissions |
404 | Resource not found |
422 | Validation error (check detail for field-level errors) |
429 | Rate limited. Back off and retry. A Retry-After header and a retry_after field (seconds) are sent only when the server knows the wait. |
Rate Limits
Requests are rate limited per workspace and, for the public contact and demo forms, per client address. A limited
request returns HTTP 429 with the standard error envelope (code, category: "retryable_transient", message).
Responses do not include X-RateLimit-* headers.
A 429 can also mean warehouse capacity is busy: creating or testing a connection, checking its readiness, or
asking a question (POST /query). When the server knows how long to wait (for example, when warehouse capacity
is busy), the response also carries a
Retry-After header and a retry_after field, both in seconds. Browser code on another origin cannot read the header;
use the retry_after field. When a request limit is exceeded, neither is sent: back off exponentially and retry.
Limits vary by plan. Contact your account team for plan-specific quotas.
Enterprise endpoints
SCIM provisioning, audit log access, billing, and analytics endpoints are documented separately and are available to enterprise customers. Contact your account team for the enterprise API guide.