Developer platform
Context through an API.
Use the TypeScript SDK, REST or MCP to connect applications and AI tools to retained knowledge.
01
A scoped client for user data.
The SDK package is @enkeon/sdk. Its exported factory remains thinkingroot for compatibility. Keep project keys on the server and scope user data to an authenticated identifier.
import { thinkingroot } from "@enkeon/sdk";
const client = thinkingroot({
gatewayUrl: "https://api.enkeon.com",
projectKey: process.env.ENKEON_API_KEY!,
});
const user = client.scope(authenticatedUserId);
await user.store([{ statement: "The pilot starts October 8." }]);
const context = await user.recall("Pilot start date?");02
Choose the integration boundary.
Use MCP for assistant tools, the SDK for typed application calls, and the API reference for HTTP endpoints and payloads.
MCP
Connect compatible assistants to project knowledge.
TypeScript
Use scoped clients and the package types to build an integration.
REST
Consult the published OpenAPI reference for the gateway surface.
03
Validate the complete workflow.
Test retrieval against facts you supplied. Check user boundaries, revision behavior, errors and empty results. Benchmark with your own model and workload before extrapolating study results.
- Protect server credentials
- Inspect source and scope
- Handle missing context
- Measure latency at the application boundary
Keep exploring
See what context makes possible.