Lamatic/AgentKit

AgentKit – Open‑source SDK for building AI agents

What it is – A framework (the Lamatic SDK) that lets you assemble, run and deploy agentic workflows with a few lines of configuration. The repo ships a registry of 70 ready‑made “kits” – complete Next.js apps, multi‑step pipelines (bundles) or single‑step templates – that cover common business‑oriented use‑cases such as support triage, code review, RAG‑powered chat, hiring automation, legal or medical assistants, and more.

How it works – Each kit lives under kits/ and contains a lamatic.config.ts that declares its type (kit, bundle or template). The SDK provides:

• A visual flow editor (the “Collaborative Studio”) where you can connect nodes like LLM calls, vector search, web‑scraping, document parsing, Google Sheets, Firecrawl, etc.
• Server‑less deployment – kits are packaged as Next.js apps that can be deployed to Vercel or any Node host with a single command.
• Enterprise‑grade security and a pluggable architecture for custom data sources.

What you get out of the box

Category	Example kits	What they demonstrate
Full apps (kits)	Support‑triage, Founder‑Lens, Hiring Copilot, Embedded Chat	End‑to‑end UI + backend, ready to run on Vercel.
Multi‑flow bundles	Document‑Parsing, GitHub‑Manager, Knowledge‑Chatbot	Re‑usable pipelines that can be invoked from other flows.
Single‑step templates	Article Summariser, Currency Converter, Image‑Based Product Identification	Minimal examples showing how to call an LLM, fetch data, or do vector search.

Typical workflow to create your own agent

1. Pick a starter kit (or copy a template).
2. Edit lamatic.config.ts to set your API keys, data sources, and any custom prompts.
3. Use the Studio UI to drag‑and‑drop nodes, add branching logic, or chain other kits.
4. Run npm run dev locally; when satisfied, deploy with npm run deploy (Vercel) or any server‑less platform.

Why it matters – Building reliable, production‑ready AI agents usually requires stitching together many services (LLMs, vector stores, web crawlers, authentication, UI). AgentKit abstracts that plumbing into reusable, declarative flows, letting teams focus on agent logic rather than infrastructure. The large collection of community‑contributed kits also serves as a learning library for best‑practice patterns (RAG, multi‑vector search, persistent memory, etc.).

Getting started

# clone the repo
git clone https://github.com/Lamatic/AgentKit.git
cd AgentKit
# install deps
npm ci
# run a sample kit, e.g. the embedded chat demo
cd kits/embed-chat
npm run dev   # opens http://localhost:3000

Documentation & support – Official docs at https://lamatic.ai/docs, a discussion forum on GitHub, and CI badges showing automated code‑review tooling.

License – The repository is open source (the README does not specify a license, so check the LICENSE file in the repo for the exact terms).

---

All details are taken directly from the repository’s README; no additional features are inferred.