• A no-code platform
• Process:
  • Assigning a goal to the agent
  • Witnessing its thinking process
  • Formulation of an execution plan
  • Taking actions accordingly
• Uses OpenAI functions
• Supports gpt-3.5-16k, pinecone and pg_vector databases
• Stack
  • Frontend: NextJS + Typescript
  • Backend: FastAPI + Python
  • DB: MySQL through docker with the option of running SQLite locally

• Documentation
• Website
• GitHub

• An LLM-powered autonomous agent platform
• A framework for autonomous agents who can work together to accomplish tasks
• Interaction with agents done via console direct messages

• Author: eumemic
• Website
• GitHub
• Twitter

• Chains together LLM "thoughts", to autonomously achieve whatever goal you set
• Internet access for searches and information gathering
• Long-term and short-term memory management
• Can execute many commands such as Google Search, browse websites, write to files, and execute Python files and much more
• GPT-4 instances for text generation
• Access to popular websites and platforms
• File storage and summarization with GPT-3.5
• Extensibility with Plugins
• "A lot like BabyAGI combined with LangChain tools"
• Features added in release 0.4.0
  • File reading
  • Commands customization
  • Enhanced testing

• Twitter
• GitHub
• Facebook
• Linkedin
• Discord
• Author: Matt Pogla

• Creates tasks based on the result of previous tasks and a predefined objective.
• The script then uses OpenAI's NLP capabilities to create new tasks based on the objective
• Leverages OpenAI's GPT-4, pinecone vector search, and LangChainAI framework
• Default model is OpenAI GPT3-turbo
• The system maintains a task list for managing and prioritizing tasks
• It autonomously creates new tasks based on completed results and reprioritizes the task list accordingly, showcasing the adaptability of AI-powered language models

• Paper: Task-driven Autonomous Agent Utilizing GPT-4, Pinecone, and LangChain for Diverse Applications
• Founder's twitter
• Twitter thread describing the system

• Just 300 lines of code
• This was built as a continued iteration on the original BabyAGI code in a lightweight way. Differences to BabyAGI include the following:
  • Task Creation Agent runs once
  • Execution Agent loops through tasks
  • Task dependencies for pulling relevant results
  • Two tools: search_tool and text_completion
  • “Mini-agent” as tool
  • Search tool combines search, scrape, chunking, and extraction.
  • Results combined to create summary_report

• Twitter
• GitHub
• Author: @yoheinakajima (Twitter)

• Powered by GPT-4 and semantic code search, precise code navigation
• Built on stack graphs and scope queries
• Fast code search and regex matching engine written in Rust
• Allows to find Code on Rust and Typescript
• Allows to stage changes

1. AI user agent: give instructions to the AI assistant with the goal of completing the task.
2. AI assistant agent: follow AI user’s instructions and respond with solutions to the task

• Paper - CAMEL: Communicative Agents for “Mind” Exploration of Large Scale Language Model Society
• Colab demo
• GitHub
• Hugging face datasets
• Slack
• Authors: Guohao Li∗ Hasan Abed Al Kader Hammoud* Hani Itani* Dmitrii Khizbullin, Bernard Ghanem

• One of the easiest and fastest no-code platform I have encountered
• After creating new agent, picking a model, data and other settings, they are ready to be deployed to website, Slack, Crisp, or Zapier
• Limit of agent in the free version
• Stack
  • Next.js
  • Joy UI
  • LangchainJS
  • PostgreSQL
  • Prisma
  • Qdrant
• Features
  • Streamline customer support, onboard new team members, and more
  • Load data from anywhere
  • No-code: User-friendly interface to manage your datastores and chat with your data
  • Securized API endpoint for querying your data
  • Auto sync data sources (coming soon)
  • Auto generates a ChatGPT Plugin for each datastore

• Documentation
• Discord
• GitHub

• GitWit uses a GPT-based agent to generate code and git to track changes made to files
• GitWit ties together large language models and modern developer tools
• It can spawn and modify codebases using just a single prompt
• GitWit is primarily aimed at full-stack developers, and is particularly loved by those with a learning mindset—such as those learning a new stack or technology
• It is in early beta and may require some experimentation with the prompts you enter
• You are offered to choose from code bases:
  • React + NextJS
  • Python using pip
  • A Chrome extension written in JavaScript
  • An AngularJS using npm.
  • Custom stack

• Model: GPT 4
• Specify your project, and the AI agent acks for clarification, and then constructs the entire code base
• Features
  • Made to be easy to adapt, extend, and make your agent learn how you want your code to look. It generates an entire codebase based on a prompt
  • You can specify the "identity" of the AI agent by editing the files in the identity folder
  • Editing the identity, and evolving the main_prompt, is currently how you make the agent remember things between projects
  • Each step in steps.py will have its communication history with GPT4 stored in the logs folder, and can be rerun with scripts/rerun_edited_message_logs.py

• Author: Anton Osika
• Twitter review by @Attack

• Lindy is still in a beta version
• Features
  • Lindy triages your email
  • She learns from your inbox and automatically surfaces the highest-priority emails for you
  • Automatic conflict handling
  • Daily briefing
  • Contract management
  • Meeting note taking
  • Summarization

• Languages: Python
• Default model: GPT-3.5-turbo (also possible with GPT-4)
• Modular Auto-GPT Framework
• Plug N Play" API - Extensible and modular "Pythonic" framework, not just a command line tool
• Features
  • "Easy to add new features, integrations and custom agent capabilities, all from python code, no nasty config files!"
  • "Minimal prompt overhead - Every token counts. We are continuously working on getting the best results with the least possible number of tokens."
  • "Human in the Loop - Ability to "course correct" agents who go astray via human feedback."
  • "Full state serialization - can save the complete state of an agent, including memory and the states of its tools to a file or python object. No external databases or vector stores required (but they are still supported)!"

• Most of the description on readme is inspired by the original privateGPT
• Model: Vicuna-7B
• Using InstructorEmbeddings
• Both Embeddings as well as LLM will run on GPU. It also has CPU support if you do not have a GPU
• Built with Langchain

• YouTube - LocalGPT: OFFLINE CHAT FOR YOUR FILES [Installation & Code Walkthrough]

• MiniAGI is a minimal general-purpose autonomous agent based on GPT-3.5 / GPT-4
• Can analyze stock prices, perform network security tests, create art, and order pizza
• MiniAGI is a simple autonomous agent compatible with GPT-3.5-Turbo and GPT-4. It combines a robust prompt with a minimal set of tools, chain-of-thoughts, and short-term memory with summarization
• It is also capable of inner monologue and self-criticism

• "Whether you are a technical person or a stakeholder, you can use Pezzo effectively. We don't believe that AI prompts should be designed in a developer's code editor. Aside from teh technical issues with this approach, it blocks productivity."
• Features
  • Centralized Prompt Management: Manage all AI prompts in one place for maximum visibility and efficiency.
  • Streamlined Prompt Design, Publishing & Versioning: Create, edit, test and publish prompts with ease.
  • Observability: Access detailed prompt execution history, stats and metrics (duration, prompt cost, completion cost, etc.) for better insights.
  • Troubleshooting: Effortlessly resolve issues with your prompts. Time travel to retroactively fine-tune failed prompts and commit the fix instantly.
  • Cost Transparency: Gain comprehensive cost transparency across all prompts and AI models.
  • Simplified Integration: Reduce code overhead by 90% by consuming your AI prompts using the Pezzo Client, regardless of the model provider.

• Documentation
• GitHub

• Human-centric, coherent whole program synthesis

• Your own junior developer

• Allows to develop, debug, and decompile

• 200 LOC, half english

• 100k context can summarize both content and codebases

• Markdown is the best prompting DSL

• Copy and paste your errors as prompts

• Copy and paste curl output as prompts

• Write CSS animation by describing what u want

• GPT4 >>> GPT3.5/Anthropic Claude for codegen

• Links

• Author: Swyx

• Demo

• Twitter

• Meme

• GitHub
• Documentation
• Discord

• Model: GPT-4
• Language: Python
• Uses OpenAI and Pinecone to give memory to an AI agent and also allows it to "think" before making an action (outputting text)
• Also, just by shutting down the AI, it doesn't forget its memories since it lives on Pinecone and its memory_counter saves the index that it's on
• A process that happens every time the AI is queried by the user:**
  • AI vectorizes the query and stores it in a Pinecone Vector Database
  • AI looks inside its memory and finds memories and past queries that are relevant to the current query
  • AI thinks about what action to take
  • AI stores the thought from Step 3
  • Based on the thought from Step 3 and relevant memories from Step 2, AI generates an output
  • AI stores the current query and its answer in its Pinecone vector database memory

• Created by @sean_pixel !

• Researchers from Stanford and Google created an interactive sandbox env with 25 Gen AI agents can simulate human behavior
• They walk in the park, join for coffee at a cafe, and share news with colleagues. They demonstrated surprisingly good social
• Westworld's inspiration is drawn from Unity software and Unity ML Agents, adapted in Python
• Languages
  • The library is available on PyPi via pip install westworld
  • Javascript version (being developed)
• Features
  • Easy creation of Grid and non-grid environments
  • Objects (Agents, Obstacles, Collectibles, Triggers)
  • Subclassing of different objects to create custom objects
  • Spawner to generate objects randomly in the environment
  • Basic rigid body system for all objects
  • Simple agent behaviors (pathfinding, wandering, random walk, fleeing, vision range)
  • Automatic maze generation
  • Layer integration to convert image to obstacle and snap it to a grid
  • Sample simulations and sample agents for classic simulations
  • Simulation visualization, replay and export (gif or video)

• GitHub
• Documentation Underlying paper - Generative Agents
• A paper simulating interactions between tens of agents
• Presenting an architecture that extends a language model to store and synthesize the agent's experiences, enabling dynamic behavior planning in an interactive sandbox environment with generative agents

• WorkGPT is an agent framework in a similar fashion to AutoGPT or LangChain. You give it a directive and an array of APIs and it will converse back and forth with the AI until its directive is complete.
• For example, a directive could be to research the web for something, to crawl a website, or to order you an Uber. We support any and all APIs that can be represented with an OpenAPI file.
• WorkGPT now has OpenAI's new function invocation feature baked into it
  • While chaining together APIs was possible before (see AutoGPT), it was slow, expensive, and error prone
  • The tweet announcing this feature

• Author: Alex MacCaw

• "This is a Swift port of BabyAGI, an example of an AI-powered task management system that uses OpenAI and Pinecone APIs to create, prioritize, and execute tasks. The main idea behind this system is that it creates tasks based on the result of previous tasks and a predefined objective."

• Author: PJ Gray

• A developer-centric approach to building internal UIs and workflows
• Turning APIs, SQL queries, and scripts into apps for the entire team
• Features
  • Airplane lets you turn SQL queries, JavaScript/Python code, HTTP requests, etc into tasks
  • Allows to run tasks through a no-code dashboard
  • Tasks for customer support, on-call runbooks, and scheduled tasks

• Profile of the company
• Documentation
• Twitter
• [They're building an AI assistant here[(https://www.airplane.dev/autopilot)

• Breaks down a high level research question into a step-by-step plan, and executes it
• Diverse tools, including a full web browser
• Can access internet information without the need for an API
• "We don't generate content using AI, as it can be unreliable. Instead, we extract relevant information from trusted sources, cluster and process it into a user-friendly format."
• AI-powered query planner intelligently routes and executes requests, ensuring correctness and diverse source selection

-Discord

• Create an Issue
• Approve the Implementation Plan
• Review the Pull Request
• Features
  • Writing code
  • Reviewing changes
  • Addressing comments
  • Answering questions
• BitBuilder Junior
  • BitBuilder Junior is an intern-level code generation tool that lives in your GitHub repository. It follows instructions to implement unamibigous code changes by changing multiple files, writing tests for generated code, and matching the style, framework, and libraries you're already using.

• GitHub Marketplace
• Slack
• [Installation(https://docs.bitbuilder.ai/install)]

• No coding required thanks to user-friendly interface
• Full SEO optimization
• Picture Upload: Users can conveniently upload and regenerate their own pictures for unlimited customization of their profiles
• Profile Customization: Users have the flexibility to customize their profiles by hiding sections, adding social media links, and sharing contact details, allowing them to showcase their unique personality and brand
• Instant Preview: Users can instantly visualize their profile changes through a conveniently placed preview button, ensuring a quick assessment of the desired appearance
• 30% Faster Speed: The app achieves an impressive 30% increase in website generation speed, providing users with a fast and efficient website building experience.

• Friendly UI for building AI agents

• Author: Sully Omarr

• Magic Copy writes, edits, and rewrites Figma text layers so you can design with real copy
• Generating images in Figma while designing
• Magic Rename intelligently names your layers so you can spend more time designing
• Magician works right inside your favorite design tool (e.g., Figma)
• Possible to get all the latest AI design advancements + future spells in one convenient plugin

• In aplha stage
• It’s not supposed to be just another coding copilots like GitHub Copilot or Codeium, but autohomous agents capable of autonomously building software from end to end

• CEO

• Building and managing LLM powered applications
• A cloud-based platform-as-a-service that allows developers to build smart agents that couple LLMs with back-end logic to interface to data, systems, and tools

• GitHub
• Fixie Developer Portal
• AI.JSX
• Twitter
• Discord

• Provides a core prompt for building the foundation of your application
• Allows you to add unlimited sections, each of which is a prompt representing a specific part of your app
• Features
  • Friday utilizes GPT-4 for AI assistance, but it has been tested and optimized with GPT-4-32k for improved speed and better results.
  • It requires 2 small requests for your app's base and 1 request per section you provide.
  • Friday employs esbuild behind the scenes for every app created by it.

• Author: Amirreza Salimi

• Grit uses machine learning and static analysis to auto-generate pull requests for cleaning up technical debt
• Users can declare how they want their code to be structured and let Grit rewrite it for them

• Linkedin
• Twitter

• "A suite of powerful AI features meant to augment data people"
• Hex can explain and document your code
• Hex Magic features know about database schemas, past operations, and the project’s execution graph, so they can make deeper, more insightful recommendations

You can see more – and sign up for the waitlist – over here.

• Launch post

• Personal assistant for life: to keep you on top of your calendar, tasks and information. Was at Llama event demo

• Twitter
• Author: Alex Graveley

• The agent runs and controls the local Google Chrome, which allows it to interact with the world/services/web apps, just like people interact with the world/services/web apps using Google Chrome
• The agent itself probably also runs locally and currently, it needs the local Google Chrome to function
• Our understanding from the demo video is that they use local code and a custom plugin in ChatGPT to control a web browser (e.g., Google Chrome). This setup enables MultiOn to perform tasks like ordering plane tickets as if a human were interacting with the browser directly
• Use cases
  • A lot of cool real use cases, e.g., -Sending an email fully autonomously -Posting a tweet -Sending a tweet reply to a specific person with a specific message -Sending a Facebook message to a friend -Searching for vacation rentals and check pricing for an upcoming trip -Searching for a wedding venue and starting the wedding planning process -Scheduling a car wash
• After introducing the GPT function calling, MultiOn can call itself recursively to spawn more sub-agents
• Instead of calling multiple functions or APIs you just need one Universal Function that can interact with all services and have it call itself to accomplish more complex tasks in parallel

• Twitter
• Author: Div Garg

"Build your team of AI agents that work for you. Early access now live. Join waitlist."

• Twitter

• Generating content
• Brainstorming ideas
• Translation
• Grammer check
• Roadmap

• Twitter
• Slack
• Discord

• Migrate frameworks such as Angular to React, libraries such as Redux to React Context, or languages such as JavaScript to TypeScript
• erform major version upgrades on any number of applications, of any size
• Upgrade frameworks such as Next.js 12 to 13, libraries such as MUI 4 to 5, or languages such as Python 2 to 3
• Roadmap

• YCombinator
• Twitter
• Linkedin¨
• Discord
• Founder: Eric Rowell

• To install, click the install button
• Then add the repository you want, make a quick ticket (e.g. writing tests)
• Prepend the ticket with "Sweep:" and let Sweep handle the rest

• GitHub
• Discord
• Tricks for prompting Sweep