Giter Club home page Giter Club logo

Missing Building Blocks in the Agent World

"AI Agents will likely be the most impactful technology of our generation". In this document I present you the following modular building blocks, fully open source. Beware that most of it is still a work in progress! See this high level overview for the current status.

Agent OpenAPI

Turn any API into an Agent, Turn any Agent into an API.

The Agent OpenAPI serves an OpenAPI for talking to an agent, so it can be discovered publicly, and can be used as a tool for other agents. More information in the Agent OpenAPI GitHub

Agent Relay

Agents need to be accessible from anywhere. The Agent Relay makes agents accessible from messaging apps, VoIP and phonecalls, and over email! Check out the Agent Relay on GitHub

CRUD OpenAPI

Data needs to be discoverable as tools. A reliable CRUD Agent is extremely useful. More info here

Why

  • We're living through a technological paradigm shift that will change how we interact with computers, and how humans can find purpose. A new foundation is being created now. In this important time, I want to do my part setting good standards for HMC that benefits humanity.

  • Big tech capitalism is trying to create a controlled closed ecosystem for AI. As AGI is approaching, misaligned commercial incentives become ever more extreme, and I don't want to live in this walled garden distopia. The solution is an open, accessible, modular ecosystem for AI Agents. An ecosystem without any vendor lock-in or privacy problems. An ecosystem where we, the people, stay in control.

Highlevel ActionSchema

This is the current ecosystem of projects developed by Code From Anywhere (❗️ dependency, ⏸️ paused, 🚫 blocked, πŸ”΄ not started, 🟠 work in progress, 🟒 done)

Name Purpose Status MVP LOC
CRUD OpenAPI Turn database into agent-tools 🟩🟩🟩🟩πŸŸ₯ 🟒 CRUD Only firsst
🟒 Semantic search
🟒 CLI
🟒 CRUD-Agent
πŸ”΄ Config: user separation
Β±3k
Agent OpenAPI Turn any API into an Agent 🟩🟩🟧πŸŸ₯πŸŸ₯πŸŸ₯ 🟒 Simple POC
🟒 OpenAPI-centric Refactor
🟠 Use tools from OpenAPIs with OAuth2
πŸ”΄ Agent Creator Agent
πŸ”΄ Files
πŸ”΄ Threads
Β±2k
Agent Relay Make agent available anywhere 🟩🟩🟩πŸŸ₯πŸŸ₯πŸŸ₯πŸŸ₯ 🟒 Browser & Phonecall STS
🟒 Custom agent compatibility
🟒 Whatsapp, SMS, Messenger
πŸ”΄ Agent-first refactor
πŸ”΄ Email
πŸ”΄ Deepgram STS Tool use
πŸ”΄ Outbound
1175
8175

A dependency to the above is what I call "OpenAPI-first development". It is an opinionated way of design-first development where your OpenAPI serves as the SSOT for a lot of things, and you don't generate it, you rather generate pieces in your code FROM it.

Strategy

ActionSchema for Devs: OEF

  • Devs want Open source. Give it.
  • Devs want Easy: Serve it BYOK, accessible, and useful.
  • Devs want Freedom. Provide them agents so they can go Screenless.

Key insights

  • Most AI is focused around realtime co-pilots because we're all still used to the direct HMC. Try making ambient pilots that don't need to be fast.
  • Pick my focus. Big topics like browser automation APIs and video editing are done by hundreds of companies and are extremely hard to stay competitive in; It's a never-ending cat and mouse game.
  • Products and APIs change all the time. Instead of choosing to spend knowledgework time in specific niches, index all available capabilities.
  • Most users care about their privacy and would want to have things ran locally. However, running locally is hard to setup and scale. Another way to have practical privacy is to keep the core local, but run smaller fleeting tasks in the cloud.
  • How any API works exactly doesn't need to be abstracted away from. The only thing we need to do is determine API capability, quality, speed, cost, and availability.

Let's Code From Anywhere!

Welcome to Code From Anywhere - a group of distributed developers and entrepreneurs building planet-first & humane-centered software. We work remotely but often come together in places like Nepal and Brazil, going on adventures.

We 🀍 Developers, AI Startups & Adventurers. Do you have a question, comment, or want to connect? Head over to our Discord

License

License: MIT

This project is licensed under the MIT License - see the LICENSE file for details.

Commercial License

If your company generates more than $1,000,000 in Annual Recurring Revenue (ARR), you are required to obtain a commercial license. Please see the COMMERCIAL_LICENSE file for more information.

Contact

For commercial licensing inquiries, please contact Wijnand at wijnand AT karsens DOT com

Code From Anywhere's Projects

actionschema icon actionschema

A declarative language allowing for pseudo-deterministic JSON Document Generation.

actionschema-migrate icon actionschema-migrate

This CLI performs a migration on remote ActionSchema microservices and creates an SDK and Typescript Types in your codebase afterwards

agent-relay icon agent-relay

Make agents accessible from messaging apps, phonecalls, and over email!

from-anywhere icon from-anywhere

Useful utilities for JS, TS, React, Node, Bun, Next, and more.

jsapi icon jsapi

This api exposes .js files over CDN

learn_prompting icon learn_prompting

Prompt Engineering, Generative AI, and LLM Guide by Learn Prompting | Join our discord for the largest Prompt Engineering learning community

marp-api icon marp-api

Turn MARP code into a HTML presentation

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    πŸ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. πŸ“ŠπŸ“ˆπŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❀️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.