Skip to main content

MCP Tools & Servers API

The MCP Tools & Servers API provides programmatic access to Nexla's MCP-as-a-Service platform. Create AI tools from Nexsets, organize them into ToolSets, deploy MCP servers, register external MCP servers via Gateway, execute tools directly, and query execution receipts.

Base URL​

All API requests are made against:

https://api-genai.nexla.io/

All endpoints are prefixed with /v1/ unless otherwise noted.

Authentication​

All API requests require authentication via a Nexla service key passed in the Authorization header:

Authorization: Bearer <NEXLA_SERVICE_KEY>

All resources are scoped to the authenticated user's organization. For MCP endpoints, OAuth (Google and Azure) is also supported. See Authentication for details.

API Reference​

Endpoint GroupBase PathDescription
Tools/v1/toolsCreate, manage, and execute AI tools
ToolSets/v1/toolsetsOrganize tools into deployable bundles
MCP Protocol/mcp/*MCP Streamable HTTP and session management
Gateway/v1/toolsets/{id}/gatewayRegister external MCP servers and sync tools
Receipts/v1/receiptsQuery execution audit trail

Quick Start​

Create a ToolSet from existing Nexsets with a single request:

curl -X POST https://api-genai.nexla.io/v1/toolsets:from_nexsets \
  -H "Authorization: Bearer $NEXLA_SERVICE_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "My AI ToolSet",
    "nexset_ids": [123, 456],
    "create_export": true
  }'

The response includes the newly created ToolSet with its auto-generated tools and an MCP server URL you can point any MCP-compatible client to.

Common Patterns​

Pagination​

List endpoints accept limit and offset query parameters. The maximum page size is 200.

GET /v1/tools?limit=50&offset=100

Filtering​

Most list endpoints support filtering by status, search, tags, and date range parameters to narrow results.

Error Responses​

Errors follow a standard JSON format:

{
  "error_code": "RESOURCE_NOT_FOUND",
  "message": "ToolSet with id 999 not found",
  "details": {}
}

Rate Limiting​

The API enforces per-principal rate limits. When exceeded, requests receive an HTTP 429 response with a Retry-After header. See Rate Limiting & Error Handling for thresholds and best practices.