🎉 Hey folks, welcome to the https://agentgpt.reworkd.ai! Discord server! 🎉 This server is designed for anyone interested in the project, whether you have questions, need assistance, or want to discuss potential issues. Feel free to share your thoughts, feedback, or simply chat with like-minded individuals passionate about the world of autonomous AI agents. I, @Adam and @swijiislit have a lot of ambitious goals for AgentGPT which we'll be sharing here. We're really excited for you all to join us on this journey 🚀

Hey all, sorry for the with AgentGPT flakiness! After launching AgentGPT, the amount of users coming as far exceeded anything we initially expected. While such a strong interest in AgentGPT super exciting, it does pose some scalability problems for us (mainly with hosting costs). After just two days, we have already exceeded the vercel limits associated with the hobby plan. We are upgrading to the "pro" plan, but that should only give us another 5 days most before we use up all 1000gb/hrs included for serverless functions. We are hoping to mitigate some costs by letting users use their own keys (doing so would allow us to skip the serverless functions). Any other ideas on cost mitigation would be very much welcome 😅

Sorry for the brief outage, it seems our OpenAI key has been exhausted, We are actively working getting AgentGPT back online 🔐

We're back online!

Hey all! Just wanted to share with you that AgentGpt is now in beta 2.0. Huge shout out to all of the community members that made this possible! 🎉 🎉 🎉 In a little over a week we have surpassed 750k+ users, 16k+ stars, and a discord community 10k+ strong. Due to the insane reception of AgentGPT we had to focus the majority of this week on scalability. We are currently averaging over 1.5M+ tasks executed daily! ✨ Highlights - Ability to add your own API key - Agents can now run on edge and directly in the browser using WASM - Ability to save and share your agents with others - Ability to export agent results to pdf, image, or text - GPT-4 support if you supply your own api key - Improvement to general UI/UX - ...and much more! Over the next couple weeks, we will be now focusing on making your agents more powerful! Web search and contextual actions are coming soon! Release notes: https://github.com/reworkd/AgentGPT/releases/tag/v.0.2.0-beta

Hey folks, we've recently updated our documentation to shine more light on the project, its roadmap, and its FAQ. Find it here: https://reworkd.github.io/AgentGPT-Documentation/docs/intro Note that it's still a WIP, if you want to see anything documented or would like to add documentation, please DM me!

Hey all! Just wanted share that AgentGpt just hit beta 3.0. Huge shout out to all of you that made it possible! 🥳 Beta 3.0 Highlights -------------------- * 🔍 Web search (Still in preview), AgentGPT now has the ability to search google and summarize results! * 🌐 Full internationalization (i18n) support, AgentGPT now natively close to 20 languages! * ⏸️ Ability to pause and resume agents, a highly requested feature. * 💄 Improved output formatting * 🎨 Improvement to general UI/UX * 📖 Improved docs and roadmap * 🐛 Various bugfixes * ...and much more! If you'd like to help shape the future of AgentGPT please fill this brief [survey](https://forms.gle/fjCo72hiYWZydJa1A). We would love to hear your feedback. ❤️ Release notes: https://github.com/reworkd/AgentGPT/releases/tag/v.0.3.0-beta

Hey all! Just an FYI we've started putting together some docs that can be found here https://docs.reworkd.ai/. If feel that there is something missing please do let us know or feel free to create a PR and/or github issue! Over the next couple weeks we are hoping to really improve the documentation for everyone 😃

Hey @everyone! Happy to announce AgentGPT just hit beta 0.4.0. Huge shout out to all of our supporters 🥳 Highlights ✨ -------------------- * 💻 Infrastructure updates in the backend. Not visible to end users but this will allow us to move faster * 🐍 Migration to a Python backend and tools support (Not yet enabled on the site but developers are free to play around with this) * 🌐 Translation updates * 🔍 Re-enabled basic web search * 📖 Additional developer docs and support for docker * 🐦 The launch of our Twitter! https://twitter.com/reworkdai What's next ❓ -------------------- * 📰 We'll be doing weekly updates about the state of AgentGPT * 🐍 Fully migrating to a Python backend * 🛠 Launching the early stage of tool integrations such as google docs AgentGPT still has a lot to improve on but we'd love to hear your feedback ❤️ If you'd like to help shape our future, please fill this brief survey https://forms.gle/fjCo72hiYWZydJa1A

Hey @everyone! We're thrilled to announce that AgentGPT just hit beta 0.5.0. Huge shout out to all of our supporters, subscribers and awesome @User team for making this possible 🥳 Now that @swijiislit @asim and I are working at this full time, we are also hoping to be more engaged with the community ~ If you ever have any question / concerns / feedback please feel free to ping us ❤️ We've also launched a "pro" subscription with access to * GPT-4* to help us finally cover our OpenAI bill and propel the development of AgentGPT forward 🚀 Use code ✨ DISCORD25 ✨ for 25% off! (Only available for the first 100 users). Sign up using the link : https://reworkd.ai/ Highlights ✨ -------------------- 💻 Full migration to our new python backend! 🎨 Image generation via Replicate 🌉 Multi-Modal tool support (web-search, images, code generation, etc..) 🚚 Moving our HQ to San Franscisco 🐦 The launch of our Twitter! https://twitter.com/reworkdai What's next ❓ -------------------- 📰 We'll be doing weekly updates about the state of AgentGPT 🛠 More advanced tooling such as google docs, github, etc.. See the full changes here: https://github.com/reworkd/AgentGPT/releases/tag/v.0.5.0-beta

Hey @everyone! Welcome to the Beta 0.6.0 release of AgentGPT! 🤩 For context, this past week, we faced some unexpected challenges, particularly in managing subscriptions and increasing latencies. We also spent considerable time refactoring. Despite these hurdles, we're excited to share some major updates. ⚡** Today, we introduced LLM Streaming! **⚡ This feature addresses the issue of high GPT API latencies, which occasionally reach 60 seconds and exceed the timeout window of our load balancer. This is a significant improvement that ensures our agents operate smoothly and efficiently. We'd like to acknowledge the Lanarky package for playing an instrumental role in this development: https://github.com/ajndkr/lanarky. 🏎️ We have more work to do regarding speed optimization. The image shared below provides a glimpse into the improvements we've made over the past week, particularly concerning our time to first byte. 🥳 As remaining open source is a key value to @asim and I, this week we also launched brand new developer cli that should help everyone get AgentGPT setup locally more easily! To get started simply fetch the repo, run ./setup.sh, and you'll have a nice walkthrough of what to do 🚀 Last week, we also launched subscriptions to help cover costs and support the further development of AgentGPT. We've received exceptional feedback, and we want to thank all our users and subscribers. The launch presented some challenges, but your continued patience and support has meant the world to us. ❤️ -Adam

Hey @everyone, it's that time again! We are excited to introduce vector databases with the 0.7.0 release of AgentGPT! ⚡️ AI agents supercharged with your data ⚡️ Big thanks to @langchainai to help enable this!. Currently we are using @weaviate by default but exploring many other #opensource providers such as @chroma-core and @supabase. It's still in its early stages but we're currently using Vector search to filter out repetitive tasks! Beyond that, this week has been mostly exploratory. We have a lot planned out now and its all public on our new roadmap page: https://github.com/orgs/reworkd/projects/3 * ✨ Highlights* - Introduced vector databases to help filter out repetitive tasks 🚀 - Improvements to UI. We are working on revamping the entire project to provide users with a more powerful interface to interact with Agents. This revamp has been heavily inspired by @tailwindcss's beautiful @tailwindui component library. 🎨 - Many bug fixes and stability improvements 🐛 See the full changes here: https://github.com/reworkd/AgentGPT/releases/tag/v.0.7.0-beta 📆 What's next? Over the next week we hope to finish up revamping our user experience. We also hope make some improvements to how AgentGPT makes use of vector databases, doing so will help to improve and personalize agents. Once again, I'd like to personally thank our amazing community for helping to make this possible! You all are awesome, we couldn't do this without you ❤️ -- Adam

@everyone We're currently experiencing issues with logging in. It seems to be stemming from issues with AWS as a whole (https://downdetector.com/status/aws-amazon-web-services/), we're working hard to resolve this. Thanks for your patience!

Hey @everyone! ! Welcome to the Beta 0.8.0 release of AgentGPT! This released focused on integrating with OpenAi's new completions functions as well as introducing support for two new vector databases 🤩 ⚡️ Massive Speed Improvements + More Vector Databases ⚡️ ✨ Highlights - Agents now use openai functions 🚨 - Vast speed improvements, we've once again HALFED our target response time (see below graph below) 🏎️ - Introduced support for two need vector databases, pg-vector and Pinecone 🍦 - Migrated settings to a distinct page to allow users to more easily configure their agents ⚙️ - Added an all new templates page to help new users get started 💪 - Agents will now provide sources / citations to their responses 📃 - As always, many small bug fixes 🐛 📆 What's next? Over the next week we hope to finish up revamping our user experience. We are also looking to start providing custom integrations to our users. We are also in the process of finally creating a landing page as we have some big announcements coming next week! I also wanted to give a special shoutout to @Json for completing his first major feature (an all new templates page)! If you'd also like to work on a feature, feel free to dm me! Lastly, you all are amazing and your support means the world to us ❤️ - Adam 🙂

Hey @everyone! We're delighted to bring you AgentGPT's version 0.9.0! We've made several enhancements with the support of our 10 new contributors, as well as @ergomake, who has significantly improved our developer experience. ✨ Major Improvements - Agents now efficiently utilize your screen with fullscreen mode 🖥️ - Control your Agents on-the-go with pause and play options ⏯️ - UI refinements, error handling upgrades, bug fixes, and more for a better Quality of Life (QOL) 🎨🚨🐛 - Featured on Decibel's open source spotlight: https://www.decibel.vc/articles/agentgpt-build-configure-and-deploy-autonomous-ai-agents-in-your-browser ⏭️ What's up next? - We're working alongside a designer to revamp our landing page and UI/UX. - Our goal is to continue enhancing user experience, exploring custom integrations, and we're gearing up for exciting announcements. - We are also doing a lot of work to make contributing easier such as creating issues in GitHub and updating out documentation Thank you everyone for your continuous support, it means everything to us ❤️ Special thanks to our new contributors: @miwgel, @chekitaderi17, @timtischler, @chennqqi, @KennyDizi, @tim-codes, @vieiralucas, @abhithory, @musabshakeel576, and @istandre. We appreciate your efforts! Stay tuned for more! 🚀 - Json 🥰

Hey @everyone! We're delighted to bring you 🎉🤖 AgentGPT 0.9.5 🤖🎉! ✨**Major Improvements** AgentGPT 0.9.5 is here, bringing you amazing new features like self-summarization and interactive chat. But that's not all – mark your calendars because in just two weeks, we have a big announcement (and a new product 🤫). Stay tuned for the big reveal and explore the latest updates in the meantime! ✨ Highlights 💬 User can now interact with agents through chat! 📕 New documentation through Mintlify! 🤯 Agents can now self-summarize with our new summary button! 🔑 Improved API key validation for custom API keys 💥 Proof of Concept (POC) for workflow automation, allowing businesses to self-automate 🤫 Lots of progress towards our secret landing page (shhhh...) ❤️ New Contributors @MichaelBergquistSuarez made their first contribution in #955 ➡️ What's up next? We have a new product cooking to supercharge automation!!!! Stay tuned 🚀 Thank you everyone for your continuous support, it means everything to us ❤️ ✅ Check it out https://agentgpt.reworkd.ai/ -Reworkd

🎉 Hey @everyone, we're super excited to release

AgentGPT 0.10.0

and announce our $1.25M pre-seed funding with support from both Y Combinator and Panache Ventures. We never expected this to happen when we first launched AgentGPT. It feels like just yesterday that we were scrambling to keep up with the overnight traffic surge that came from our release three months prior. We've learned a lot from that time: - 👻 Language models have issues with Hallucination: They make up information not grounded in any facts. - ♻️ Agents suffer from repetition: They struggle to understand the current context of a task, leading to unproductive loops. - 📉 Failure probability only compounds with longer running agents - 💼 Businesses are increasingly seeking automation through AI Agents. We're now working on a platform that solves these issues by coupling AI Agents with a structured workflow system. This enables businesses to formalize and codify their core processes in a verifiably correct way while still enabling AI Agents to reason about the nuances of a problem. If this is something of interest to you, please join our platform here: https://6h6bquxo5g1.typeform.com/to/qscfsOf1?typeform-source=reworkd.ai Thank you for all of the support thus far, we hope you continue to join us in building the future of open source AI Agent automation ❤️ Read more here: https://www.ycombinator.com/launches/J1r-reworkd-ai-the-open-source-zapier-of-ai-agents

# Discord Communities are Getting a Safety Buff 🦺 Today, we're rolling out three new features that will help you keep your communities safe from DM spam and raids: Activity Alerts, Security Actions, and a revamped Members page. ## Activity Alerts and Security Actions Activity Alerts will notify you of abnormal server behavior, particularly massive raids or unusual DM activity. From there, your team can investigate further and take further action to protect your community as necessary. If it doesn’t seem like an issue, you can easily mark it as safe. Security Actions allow swift server lockdowns by surfacing useful tools quicker, helping halt new member joins or aiding in closing off DMs between non-friend community members. Current Security Actions include: - **Pause Invites**: Stop new members from using any invites to join the server. Experienced moderators may know this one already, and now it’s more easily accessible. - **Pause DMs**: Temporarily stops new DMs from being sent between members in the server. This excludes friends DMing each other, when moderators DM server members, or DMs from Apps. ## Members Page Previously tucked away in Server Settings, the Members page is used to find specific members, display members with a specific role, and even help you prune inactive members. Now, the Members page is in a more easily-accessible spot above your server’s channel list. And it’s been reimagined, letting you view your members in a more organized manner. It now displays relevant safety information about your members: - **Join Date**: See when they joined your server. - **Invite Link**: See which invite link they used to join your server. - **Account Age**: See if they’ve been on Discord for a while or are a new user. - **Signals**: Appears if someone may have been engaging in unusual DM Activity, or flags a member you’ve timed out. Use this to easily filter and find members who might require additional action or review. Over time, we plan on surfacing additional relevant Signals. All of these new features work in tandem to provide mods with the helpful information and actions needed to help squash DM spam and thwart raids. We’re committed to improving safety on Discord by empowering and equipping mods with more tools to help protect your servers from harmful disruptions. It’s a never-ending journey, so stay tuned—the best is yet to come! For the details on how to use these new tools, check out the following articles in our help center: [Members page Help Center Article]() [Activity Alerts + Security Actions Help Center Article]() https://cdn.discordapp.com/attachments/667560445975986190/1152346782492278845/Disc_MKT_Community_Safety_Members_GIF_03_V4.mp4

🎉 Hey @everyone! It's been a while but we're excited to announce the AgentGPT 1.0 ✨ update! Featuring: - 🚀 A freshly redesigned theme - 🔎 Updated agents with a focus on information retrieval through search - 🦥 A new tool for personalized context and queries that take information from Notion/Google Drive/etc - 📈 Increased usage limits for all users - 🤖 Automatic name generation Give it a go at https://agentgpt.reworkd.ai/ and tell us what you think! Much of our time has actually been focused on a platform outside of AgentGPT. This new platform aims to use AI to take actions on websites and retrieve information. If this is interesting to you, please fill out https://6h6bquxo5g1.typeform.com/to/qscfsOf1? explaining your use case! As always, thank you for all of the support thus far! ❤️ https://cdn.discordapp.com/attachments/1094791205537202359/1169374636656578640/AgentGPT_1.0.mp4?ex=65552bff&is=6542b6ff&hm=d112600f46dd0fabff3dda43a9097407cd12078e261ddaffe29a70e3c3f02519&

Hey @everyone! We're super exited to launch launch our newest open source project, Bananalyzer! Banana-lyzer is an open source AI Agent evaluation framework and dataset for web tasks with Playwright (And has a banana theme because why not). We've created this because want to empower everyone with the tools to create high quality AI web agents. Happy Bananalyzing! 🍌🙉 https://github.com/reworkd/bananalyzer

Hey @everyone! Its been a busy week for us at the #agentgpt team! We're super excited to launch another open source project, Tarsier! If you're building web agents using slenium / playwright this library help you! 🚀 If you've tried using GPT-4(V) to automate web interactions, you've probably run into questions like: - How do you map LLM responses back into web elements? - How can you mark up a page for an LLM better understand its action space? - How do you feed a "screenshot" to a text-only LLM? At Reworkd, we found ourselves reusing the same utility libraries to solve these problems across multiple projects. Because of this we're now open-sourcing this simple utility library for multimodal web agents... Tarsier! The video below demonstrates Tarsier usage by feeding a page snapshot into a langchain agent and letting it take actions. Demo: https://twitter.com/khoomeik/status/1723432848739483976 Repo: https://github.com/reworkd/tarsier Thanks for the support everyone ❤️

Hey @everyone! We know it's been a long time since we've sent out an update but we've been super hard at work refining the future of our product ⚡ Specifically we've fully invested into building the next generation of web agents 🌎. What this means is that we're working on LLM systems that can understand and interact with any website. Access is still behind a waitlist but we've been working with many companies using this technology to extract web data across 100s to 1000s of websites 🤖 Video attached! Directionally its a bit of a shift from what we were working on before, but we're extremely excited for this new direction. We want to build the future of web interaction by creating a sort of unified API for the internet. Regardless of whether you need all of the government data across europe, or you'd like to enable users to purchase e-commerce items via an API, we want to be the solution for you. If you're interested, please check out https://reworkd.ai Also, we've just launched so please give us a retweet at https://x.com/asimdotshrestha/status/1810720478111371581 As always, thanks for the support! 💙 Stay tuned for a super cool product demo at the end of the month! https://cdn.discordapp.com/attachments/1094791205537202359/1260291484524085268/FINAL_PRODUCT_LAUNCH.mp4?ex=668ec968&is=668d77e8&hm=4da1f56cde927bd3f88f043a11671a6873df5859e78a119db5daeb591d98108a&

Hey @here over the past year or so we've been hyper focused on building web agents to scrape data at scale. We just did a launch on product hunt (https://www.producthunt.com/) and would love to get your support! https://cdn.discordapp.com/attachments/1094791205537202359/1349060990444109875/Gallery_1.png?ex=67d1ba74&is=67d068f4&hm=ab534b56a3583d94c7a124af70ce35f7d1cfaf85d9f5b187d5615d7a6517be92&

Currently at #9! https://cdn.discordapp.com/attachments/1094791205537202359/1349061053262205051/image.png?ex=67d1ba83&is=67d06903&hm=f06568fba5428cb6d3182a26fea60a5aefc63a6d76cb8ae679fa1a746e357cb3&