Public SDKs, docs, and examples for real evaluation

Knowledge² gives technical teams a public developer surface they can actually use: install the SDK that fits their stack, read the docs, run the examples, and integrate through common AI frameworks.

Platform

Public resources for serious evaluation

Start with the public Knowledge² Python SDK on GitHub.
Use the public Knowledge² TypeScript SDK for Node and Next.js backends.
Use the public docs repo for getting-started guides, integrations, and API reference.
Run the public examples repo to evaluate fit before deeper implementation work.

Knowledge² Python SDK

Official Python client with public source on GitHub
One client spanning ingest, search, tuning, deployment, and operations
A straightforward starting point for programmatic evaluation

Playground search showing scored results and source provenance

Knowledge² TypeScript SDK

Official Node and TypeScript client with public source on GitHub
Covers the core retrieval workflow plus agents, feeds, pipelines, and A2A
A clean fit for Next.js backends, worker services, and Node tooling

Support search agent configuration with collection, model, task type, and instructions

Knowledge² docs

Public getting-started guide and API reference
Framework integration documentation for LangChain and LlamaIndex
The clearest view of the hosted developer surface before implementation

API key management showing project and organization access scopes

Knowledge² examples

Runnable SDK lifecycle examples
Released LangChain and LlamaIndex examples for common adoption patterns
A practical way to test implementation fit with less guesswork

Generated answer with cited evidence and source passages

Production-friendly integration behavior

Automatic retries and rate-limit handling
Pagination and predictable error handling for long-lived integrations
Auth patterns that support both hosted and self-managed deployments

Usage dashboard with request volume, error rate, latency, and collection activity

Technical teams can test implementation fit from public repos instead of sales-only material.
The first working search flow should fit in a small script.
Teams can layer K² into LangChain or LlamaIndex when that is the right fit.

Example quickstart outcome

A product engineer builds a first search flow

The goal is a small, usable baseline that the team can evaluate and extend.

Question

How do I rotate an API key?

Returned result

API keys can be rotated from Settings > Security. Existing keys remain valid until revoked, which lets you stage the rollover safely.

Project created: Support Search
Corpus created: Product Docs
First search result returned

Published Knowledge² TypeScript SDK quickstart

Use the same surface documented in the public SDK

TypeScript SDK example

TypeScript

import { Knowledge2 } from "@knowledge2/sdk";const client = new Knowledge2({ apiKey: process.env.K2_API_KEY! });const corpus = await client.corpora.create(  "project-id",  "Product Docs",  "Support content",);await client.documents.upload(corpus!.id, {  rawText: "API keys can be rotated from Settings > Security...",  metadata: { source: "security-docs" },});await client.indexes.build(corpus!.id);const results = await client.search.query(corpus!.id, {  query: "How do I rotate an API key?",  topK: 5,});console.log(results?.results);