├── README.md └── assets ├── footer.png └── landscape-latest.png /README.md: -------------------------------------------------------------------------------- 1 |

2 | 💠 Awesome Devins 3 |

4 | 5 | 6 | 7 | 8 | 9 | 10 |

11 |

12 |

13 | Try out E2B's cloud runtime for AI agents 14 |

15 | 16 | 17 |
👉 Submit new product here
18 | 19 | Chart of AI Agents Landscape 20 | 21 | Welcome to our list of AI agents that are inspired by release of Devin. See the 🌟 [complete list of AI agents on GitHub](https://e2b.dev/ai-agents) or in [web UI](https://e2b.dev/ai-agents). 22 | 23 | The list is done according to our best knowledge, although definitely not comprehensive. Check out also the Awesome List of SDKs for AI Agents. 24 | Discussion and feedback appreciated! :heart: 25 | 26 | Have anything to add? Create a pull request or fill in this [form](https://forms.gle/UXQFCogLYrPFvfoUA). Please keep the alphabetical order and in the correct category. 27 | 28 | 32 | 33 | ## Want to use E2B with your AI product? 34 | Contact us at [hello@e2b.dev](mailto:hello@e2b.dev) or [on Discord](https://discord.gg/35NF4Y8WSE). Follow us on [X (Twitter)](https://twitter.com/e2b) 35 | 36 | We are open-source and you can get started with E2B [here](https://e2b.dev/docs?ref=awesome-sdks). 37 | 38 | 39 | 47 | 48 | 49 | # Open-source "Devins" 50 | 51 | ## [Anterion](https://github.com/MiscellaneousStuff/anterion) 52 | Open-source software engineer 53 | 54 |
55 | 56 | ![Image](https://avatars.githubusercontent.com/u/14856541?s=64&v=4) 57 | ### Category 58 | Coding, general purpose 59 | 60 | ### Description 61 | - Anterion Agent extends the capabilities of SWE-agent to plan and execute open-ended engineering tasks. 62 | - Frontend is inspired by OpenDevin 63 | 64 | 65 | ### Links 66 | - [GitHub](https://github.com/MiscellaneousStuff/anterion) 67 | - [Discord](https://discord.com/invite/nbY6njCuxh) 68 | - [YouTube demo](https://www.youtube.com/watch?v=J-KZNFVcAxU&ab_channel=Anterion) 69 | - [YouTube review](https://www.youtube.com/watch?v=wgtZbw1yEtE&ab_channel=Prepdrive) 70 | 71 | 72 |
73 | 74 | ## [AutoCodeRover](https://github.com/nus-apr/auto-code-rover) 75 | Autonomous Program Improvement 76 | 77 |
78 | 79 | ![Image](https://private-user-images.githubusercontent.com/48704330/322199117-e10c3270-442c-4673-8656-735c892dfb66.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.Y7mGAqtPaVKWShBVUOzSALad08wyJ5ZFj28_Q_HeVlY) 80 | ### Category 81 | Coding, general purpose 82 | 83 | ### Description 84 | - This agent is based on a paper "AutoCodeRover: Autonomous Program Improvement". 85 | - It combines LLMs with analysis and debugging capabilities to prioritize patch locations ultimately leading to a patch. 86 | - Resolves ~16% of issues of SWE-bench (total 2294 GitHub issues) and ~22% issues of SWE-bench lite (total 300 GitHub issues). 87 | - AutoCodeRover works in two stages: 88 | - Context retrieval: The LLM is provided with code search APIs to navigate the codebase and collect relevant context. 89 | - Patch generation: The LLM tries to write a patch, based on retrieved context. 90 | - AutoCodeRover has two unique features: 91 | - Code search APIs are Program Structure Aware. Instead of searching over files by plain string matching, AutoCodeRover searches for relevant code context (methods/classes) in the abstract syntax tree. 92 | - When a test suite is available, AutoCodeRover can take advantage of test cases to achieve an even higher repair rate, by performing statistical fault localization. 93 | 94 | ### Links 95 | - [GitHub](https://github.com/nus-apr/auto-code-rover) 96 | - [Paper](https://arxiv.org/abs/2404.05427) 97 | 98 |
99 | 100 | ## [AutoDev](https://github.com/unit-mesh/auto-dev) 101 | Open-source Devin alternative 102 | 103 |
104 | 105 | ![Image](https://github.com/unit-mesh/auto-dev/raw/master/plugin/src/main/resources/META-INF/pluginIcon.svg) 106 | 107 | ### Category 108 | Coding, general purpose 109 | 110 | ### Description 111 | - AutoDev has multilingual support 112 | - Languages support: Java, Kotlin, JavaScript/TypeScript, Rust, Python, Golang, C/C++/OC, or others… 113 | 114 | 115 | 116 | ### Links 117 | - [GitHub](https://github.com/unit-mesh/auto-dev) 118 | - [Documentation](https://ide.unitmesh.cc/) 119 | 120 |
121 | 122 | ## [Codel](https://github.com/semanser/codel?tab=readme-ov-file) 123 | Fully autonomous AI agent 124 | 125 |
126 | 127 | ![Image](https://github.com/semanser/codel/raw/main/.github/demo.png) 128 | ### Category 129 | Coding, general purpose 130 | 131 | ### Description 132 | - Fully autonomous AI Agent that can perform complicated tasks and projects using terminal, browser, and editor 133 | - Everything is running in a sandboxed Docker environment. 134 | - It can fetch latest information from the web (tutorials, docs, etc.) if needed. 135 | - Has a built-in text editor for previewing all files 136 | - All the history commands and outputs are saved in the PostgreSQL database. 137 | - The simplest way to start Codel is to use a pre-built Docker image 138 | 139 | 140 | ### Links 141 | - [GitHub](https://github.com/semanser/codel?tab=readme-ov-file) 142 | - [Discord](https://discord.com/invite/uMaGSHNjzc) 143 | 144 |
145 | 146 | ## [Devika](https://github.com/stitionai/devika) 147 | Agentic AI Software Engineer 148 | 149 |
150 | 151 | ![Image](https://github.com/stitionai/devika/raw/main/.assets/devika-screenshot.png) 152 | ### Category 153 | Coding, general purpose 154 | 155 | ### Description 156 | - Devika supports Claude 3, GPT-4, GPT-3.5, and Local LLMs via Ollama 157 | - GitHub stars: 14.4k 158 | - Devika is an Agentic AI Software Engineer that can understand high-level human instructions, break them down into steps, research relevant information, and write code to achieve the given objective. 159 | - Devika aims to be a competitive open-source alternative to Devin by Cognition AI. 160 | - Its architecture has a Code Writing Module that generates code based on the plan, research findings, and user requirements. Supports multiple programming languages. It also has a Browser Interaction Module that enables Devika to navigate websites, extract information, and interact with web elements as needed. 161 | - To enable focused research and information gathering, Devika employs keyword extraction techniques. 162 | 163 | 164 | ### Links 165 | - [GitHub](https://github.com/stitionai/devika) 166 | 167 |
168 | 169 | ## [Devon](https://github.com/entropy-research/Devon) 170 | Open-source Devin alternative 171 | 172 |
173 | 174 | ![Image](https://avatars.githubusercontent.com/u/163390653?s=200&v=4) 175 | ### Category 176 | Coding, general purpose 177 | 178 | ### Description 179 | - Open-source alternative to Devin by Entropy research 180 | - GitHub stars: 150 181 | - Devon is an open-source SWE Agent that aims to help software engineers with the development and maintenance of software. 182 | - It doesn’t provide an extensive description, but as features it mentions reliable multi file editing or use of tools such as git 183 | - Devon is built with Python 184 | 185 | 186 | ### Links 187 | - [GitHub](https://github.com/entropy-research/Devon) 188 | 189 |
190 | 191 | ## [MetaGPT](https://github.com/geekan/MetaGPT) 192 | Agent framework returning Design, Tasks, or Repo 193 | 194 |
195 | 196 | ![image](https://github.com/geekan/MetaGPT/raw/main/docs/resources/MetaGPT-new-log.png) 197 | 198 | ### Category 199 | Multi-agent, Coding, Build your own 200 | 201 | ### Description 202 | MetaGPT is a multi-agent framework 203 | MetaGPT has existed longer than other Devin alternatives, but after the release of Devin, the MetaGPT team aims to replace most of the Devin capabilities. 204 | MetaGPT allows the assignment of different roles to GPTs to form a collaborative software entity for complex tasks. 205 | You can use this multi-agent framework in your own scenario to build your own application. 206 | Internally, MetaGPT includes agents with roles of product managers, architects, project managers, and engineers. 207 | It provides the entire process of a software company along with carefully orchestrated SOPs. Code = SOP(Team) is the core philosophy. 208 | 209 | 210 | ### Links 211 | - [GitHub](https://github.com/geekan/MetaGPT) 212 | - [Discord](https://discord.com/invite/4WdszVjv) 213 | - [Twitter](https://twitter.com/DeepWisdom2019) 214 | - [Paper - MetaGPT: Meta Programming for Multi-Agent Collaborative Framework](https://arxiv.org/abs/2308.00352) 215 | 216 |
217 | 218 | ## [⭐ OpenDevin](https://github.com/OpenDevin/OpenDevin) 219 | OpenDevin: Code Less, Make More 220 | 221 |
222 | 223 | ![Image](https://github.com/OpenDevin/OpenDevin/raw/main/logo.png) 224 | ### Category 225 | Coding, general purpose 226 | 227 | ### Description 228 | - The OpenDevin project aims to replicate the original Devin model. 229 | - By engaging the open-source community, we aim to tackle the challenges faced by Code LLMs in practical scenarios, producing works that significantly contribute to the community and pave the way for future advancements. 230 | - OpenDevin does code execution via Docker, but there is a work in progress with adding an option to execute code in the sandboxed environment 231 | - It has over 20.7k stars 232 | The key technologies used in the project include FastAPI, uvicorn, LiteLLM, Docker, Ruff, MyPy, LlamaIndex, and React. 233 | 234 | 235 | ### Links 236 | - [GitHub](https://github.com/OpenDevin/OpenDevin) 237 | - [Demo video](https://www.youtube.com/watch?v=g6-DpvBlTWY&ab_channel=GitWit) 238 | - [Discord](https://discord.com/invite/mBuDGRzzES) 239 | - [Slack](https://opendevin.slack.com/join/shared_invite/zt-2etftj1dd-X1fDL2PYIVpsmJZkqEYANw#/shared-invite/email) 240 | 241 |
242 | 243 | ## [Plandex](https://github.com/plandex-ai/plandex) 244 | Terminal-based AI coding engine for complex tasks 245 | 246 |
247 | 248 | ![Image](https://github.com/plandex-ai/plandex/raw/main/images/plandex-logo-dark.png) 249 | 250 | ### Category 251 | Coding, general purpose 252 | 253 | ### Description 254 | - Plandex uses long-running agents to complete tasks that span multiple files and require many steps. 255 | - It breaks up large tasks into smaller subtasks, then implements each one, continuing until it finishes the job. 256 | - It helps you churn through your backlog, work with unfamiliar technologies, get unstuck, and spend less time on the boring stuff. 257 | 258 | 259 | ### Links 260 | - [GitHub](https://github.com/plandex-ai/plandex) 261 | 262 |
263 | 264 | ## [SWE Agent](https://github.com/princeton-nlp/SWE-agent) 265 | Open-source Devin alternative 266 | 267 |
268 | 269 | ![Image](https://github.com/princeton-nlp/SWE-agent/raw/main/assets/swe-agent-banner.png) 270 | 271 | ### Category 272 | Coding, general purpose 273 | 274 | ### Description 275 | - This Devin alternative scores 12.3% on the FULL swe benchmark 276 | - SWE-agent works by interacting with a specialized terminal, which allows it to: 277 | - 🔍 Open, scroll and search through files 278 | - ✍️ Edit specific lines w/ automatic syntax check 279 | - 🧪 Write and execute tests 280 | - This custom-built interface is critical for good performance. Simply connecting an LM to a vanilla bash terminal does not work well. 281 | - ["Our key insight is that LMs require carefully designed agent-computer interfaces (similar to how humans like good UI design). E.g. When the LM messes up indentation, our editor prevents it and gives feedback."](https://x.com/jyangballin/status/1775114448513958134) 282 | - SWE-agent was released by the Princeton NLP team. 283 | - What makes SWE-agent special is that it performs almost as well as Devin on the SWE-bench. 284 | - It is important to say that the performance [varies](https://www.swebench.com/) based on the model used by the agent. 285 | - The changes and innovations in SWE-agent compared to Devin are: 286 | - The code in SWE Agent is executed locally via Docker. 287 | - It uses “Agent-Computer Interface” (ACI) - constraining the interface makes the agent easier to use for LMs. Only a few commants are allowed: run code, look for code, edit code and submit changes to GitHub. 288 | - Any code the agent writes goes through a syntax check (linter) before being submitted. If the syntax is incorrect, the agent gets feedback and is forced to redo the code. 289 | - The agent can only read 100 lines of code at a time, rather than the entire file. This makes it easier for the language model to understand the code. 290 | 291 | 292 | ### Links 293 | - [GitHub](https://github.com/princeton-nlp/SWE-agent) 294 | - [Web](https://swe-agent.com/) 295 | - [Demo](https://swe-agent.com/demo) 296 | - [Discord](https://discord.com/invite/AVEFbBn2rH) 297 | 298 | 299 |
300 | 301 |
302 | 303 | 304 | # Closed-source "Devins" 305 | 306 | ## [Devin](https://www.cognition-labs.com/introducing-devin) 307 | The first AI software engineer 308 | 309 |
310 | 311 | ![image](https://assets-global.website-files.com/65cf071d26e52092bc212f6e/65ed4622397bb038560f1ef3_cropped-p-500.png) 312 | 313 | ### Category 314 | General purpose, Coding 315 | 316 | 317 | ### Description 318 | Devin is in early phase now, but according to demo, it has the following capabilities: 319 | - Can learn how to use unfamiliar technologies. 320 | - Can build and deploy apps end to end. 321 | - Can autonomously find and fix bugs in codebases. 322 | 323 | 324 | ### Links 325 | - [Blog post](https://www.cognition-labs.com/introducing-devin) 326 | - [X (Twitter)](https://twitter.com/cognition_labs) 327 | 328 |
329 | 330 | ## [Fume](https://www.fumedev.com/) 331 | AI Software Engineer to 10x efficiency 332 | 333 |
334 | 335 | ![image](https://www.fumedev.com/images/fume_white.svg) 336 | 337 | ### Category 338 | General purpose, Coding 339 | 340 | ### Description 341 | Fume is an expert on everything your engineers need. It can help your engineers to solve any problem, or take the driver seat and automate solutions end-to-end. 342 | - Slack integration 343 | - Automated Code Changes 344 | - Step-by-Step Guidance 345 | - Private & Secure 346 | - SotA and Public 347 | 348 | ### Links 349 | - [Docs](https://docs.fumedev.com/introduction) 350 | - [Founder's X (Twitter)](https://twitter.com/aegucer) 351 |
352 | 353 | 354 | ## Want to use E2B with your AI product? 355 | Contact us at [hello@e2b.dev](mailto:hello@e2b.dev) or [on discord](https://discord.gg/35NF4Y8WSE). 356 | 357 | We are open-source and you can get started with E2B [here](https://e2b.dev/docs?ref=awesome-sdks). 358 | 359 | SDKs Repo Visual 360 | 361 | 362 | ## Join the community 363 | - Follow us on [X ](https://twitter.com/e2b) 364 | - [Hit us up on discord](https://discord.gg/35NF4Y8WSE) 365 | - Feel free to reach out to us at [hello@e2b.dev](mailto:hello@e2b.dev). 366 | -------------------------------------------------------------------------------- /assets/footer.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/e2b-dev/awesome-devins/a5f4db6de104b18cb569157db9d39ff26022635c/assets/footer.png -------------------------------------------------------------------------------- /assets/landscape-latest.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/e2b-dev/awesome-devins/a5f4db6de104b18cb569157db9d39ff26022635c/assets/landscape-latest.png --------------------------------------------------------------------------------